Merge branch '3.0' into feature/TD-14481-3.0

docs-cn/27-train-faq/01-faq.md

@@ -74,7 +74,7 @@ title: 常见问题及反馈
    检查服务器侧 TCP 端口连接是否工作：`nc -l {port}`
    检查客户端侧 TCP 端口连接是否工作：`nc {hostIP} {port}`
 
- - Windows 系统请使用 PowerShell 命令 Net-TestConnection -ComputerName {fqdn} -Port {port} 检测服务段端口是否访问
+ - Windows 系统请使用 PowerShell 命令 Test-NetConnection -ComputerName {fqdn} -Port {port} 检测服务段端口是否访问
 
 10. 也可以使用 taos 程序内嵌的网络连通检测功能，来验证服务器和客户端之间指定的端口连接是否通畅（包括 TCP 和 UDP）：[TDengine 内嵌网络检测工具使用指南](https://www.taosdata.com/blog/2020/09/08/1816.html)。
 

docs-en/14-reference/03-connector/python.mdx

@@ -199,8 +199,7 @@ The `connect()` function returns a `taos.TaosConnection` instance. In client-sid
 
 All arguments to the `connect()` function are optional keyword arguments. The following are the connection parameters specified.
 
-- - `url`： The URL of taosAdapter REST service. The default is <http://localhost:6041>.
-- `host`: The host to connect to. The default is localhost.
+- `url`： The URL of taosAdapter REST service. The default is <http://localhost:6041>.
 - `user`: TDengine user name. The default is `root`.
 - `password`: TDengine user password. The default is `taosdata`.
 - `timeout`: HTTP request timeout in seconds. The default is `socket._GLOBAL_DEFAULT_TIMEOUT`. Usually, no configuration is needed.

docs-en/27-train-faq/01-faq.md

@@ -45,7 +45,7 @@ When the client is unable to connect to the server, you can try the following wa
    Check whether a TCP port on server side is open: `nc -l {port}`
    Check whether a TCP port on client side is open: `nc {hostIP} {port}`
 
- - On Windows system `Net-TestConnection -ComputerName {fqdn} -Port {port}` on PowerShell can be used to check whether the port on server side is open for access.
+ - On Windows system `Test-NetConnection -ComputerName {fqdn} -Port {port}` on PowerShell can be used to check whether the port on server side is open for access.
 
 7.  TDengine CLI `taos` can also be used to check network, please refer to [TDengine CLI](/reference/taos-shell).
 

include/common/tcommon.h

@@ -108,6 +108,7 @@ typedef struct SColumnInfoData {
 
 typedef struct SQueryTableDataCond {
   // STimeWindow  twindow;
+  uint64_t     suid;
   int32_t      order;  // desc|asc order to iterate the data block
   int32_t      numOfCols;
   SColumnInfo* colList;

include/common/tdatablock.h

@@ -71,7 +71,7 @@ SEpSet getEpSet_s(SCorEpSet* pEpSet);
 #define colDataGetData(p1_, r_) \
   ((IS_VAR_DATA_TYPE((p1_)->info.type)) ? colDataGetVarData(p1_, r_) : colDataGetNumData(p1_, r_))
 
-#define IS_JSON_NULL(type,data) ((type) == TSDB_DATA_TYPE_JSON && *(data) == TSDB_DATA_TYPE_NULL)
+#define IS_JSON_NULL(type, data) ((type) == TSDB_DATA_TYPE_JSON && *(data) == TSDB_DATA_TYPE_NULL)
 
 static FORCE_INLINE bool colDataIsNull_s(const SColumnInfoData* pColumnInfoData, uint32_t row) {
   if (!pColumnInfoData->hasNull) {
@@ -180,7 +180,7 @@ static FORCE_INLINE void colDataAppendDouble(SColumnInfoData* pColumnInfoData, u
   *(double*)p = *(double*)v;
 }
 
-int32_t getJsonValueLen(const char *data);
+int32_t getJsonValueLen(const char* data);
 
 int32_t colDataAppend(SColumnInfoData* pColumnInfoData, uint32_t currentRow, const char* pData, bool isNull);
 int32_t colDataMergeCol(SColumnInfoData* pColumnInfoData, uint32_t numOfRow1, int32_t* capacity,
@@ -223,7 +223,8 @@ int32_t blockDataTrimFirstNRows(SSDataBlock* pBlock, size_t n);
 
 SSDataBlock* createOneDataBlock(const SSDataBlock* pDataBlock, bool copyData);
 
-void  blockCompressEncode(const SSDataBlock* pBlock, char* data, int32_t* dataLen, int32_t numOfCols, int8_t needCompress);
+void        blockCompressEncode(const SSDataBlock* pBlock, char* data, int32_t* dataLen, int32_t numOfCols,
+                                int8_t needCompress);
 const char* blockCompressDecode(SSDataBlock* pBlock, int32_t numOfCols, int32_t numOfRows, const char* pData);
 
 void blockDebugShowData(const SArray* dataBlocks, const char* flag);
@@ -231,6 +232,8 @@ void blockDebugShowData(const SArray* dataBlocks, const char* flag);
 int32_t buildSubmitReqFromDataBlock(SSubmitReq** pReq, const SArray* pDataBlocks, STSchema* pTSchema, int32_t vgId,
                                     tb_uid_t suid);
 
+char* buildCtbNameByGroupId(const char* stbName, uint64_t groupId);
+
 SSubmitReq* tdBlockToSubmit(const SArray* pBlocks, const STSchema* pSchema, bool createTb, int64_t suid,
                             const char* stbFullName, int32_t vgId);
 

include/common/tmsg.h

@@ -1315,6 +1315,31 @@ typedef struct {
 int32_t tSerializeSKillTransReq(void* buf, int32_t bufLen, SKillTransReq* pReq);
 int32_t tDeserializeSKillTransReq(void* buf, int32_t bufLen, SKillTransReq* pReq);
 
+typedef struct {
+  int32_t useless;  // useless
+} SBalanceVgroupReq;
+
+int32_t tSerializeSBalanceVgroupReq(void* buf, int32_t bufLen, SBalanceVgroupReq* pReq);
+int32_t tDeserializeSBalanceVgroupReq(void* buf, int32_t bufLen, SBalanceVgroupReq* pReq);
+
+typedef struct {
+  int32_t vgId1;
+  int32_t vgId2;
+} SMergeVgroupReq;
+
+int32_t tSerializeSMergeVgroupReq(void* buf, int32_t bufLen, SMergeVgroupReq* pReq);
+int32_t tDeserializeSMergeVgroupReq(void* buf, int32_t bufLen, SMergeVgroupReq* pReq);
+
+typedef struct {
+  int32_t vgId;
+  int32_t dnodeId1;
+  int32_t dnodeId2;
+  int32_t dnodeId3;
+} SRedistributeVgroupReq;
+
+int32_t tSerializeSRedistributeVgroupReq(void* buf, int32_t bufLen, SRedistributeVgroupReq* pReq);
+int32_t tDeserializeSRedistributeVgroupReq(void* buf, int32_t bufLen, SRedistributeVgroupReq* pReq);
+
 typedef struct {
   char user[TSDB_USER_LEN];
   char spi;

include/common/tmsgdef.h

@@ -153,6 +153,9 @@ enum {
   TD_DEF_MSG_TYPE(TDMT_MND_GRANT, "grant", NULL, NULL)
   TD_DEF_MSG_TYPE(TDMT_MND_AUTH, "auth", NULL, NULL)
   TD_DEF_MSG_TYPE(TDMT_MND_APPLY_MSG, "mnode-apply", NULL, NULL)
+  TD_DEF_MSG_TYPE(TDMT_MND_BALANCE_VGROUP, "balance-vgroup", NULL, NULL)
+  TD_DEF_MSG_TYPE(TDMT_MND_MERGE_VGROUP, "merge-vgroup", NULL, NULL)
+  TD_DEF_MSG_TYPE(TDMT_MND_REDISTRIBUTE_VGROUP, "redistribute-vgroup", NULL, NULL)
 
   TD_NEW_MSG_SEG(TDMT_VND_MSG)
   TD_DEF_MSG_TYPE(TDMT_VND_SUBMIT, "submit", SSubmitReq, SSubmitRsp)

include/common/ttokendef.h

@@ -188,10 +188,10 @@
 #define TK_KILL          170
 #define TK_CONNECTION    171
 #define TK_TRANSACTION   172
-#define TK_MERGE         173
+#define TK_BALANCE       173
 #define TK_VGROUP        174
-#define TK_REDISTRIBUTE  175
-#define TK_SPLIT         176
+#define TK_MERGE         175
+#define TK_REDISTRIBUTE  176
 #define TK_SYNCDB        177
 #define TK_DELETE        178
 #define TK_NULL          179

include/libs/nodes/cmdnodes.h

@@ -28,6 +28,14 @@ extern "C" {
 #define DESCRIBE_RESULT_TYPE_LEN  (20 + VARSTR_HEADER_SIZE)
 #define DESCRIBE_RESULT_NOTE_LEN  (8 + VARSTR_HEADER_SIZE)
 
+#define PRIVILEGE_TYPE_MASK(n) (1 << n)
+
+#define PRIVILEGE_TYPE_ALL   PRIVILEGE_TYPE_MASK(0)
+#define PRIVILEGE_TYPE_READ  PRIVILEGE_TYPE_MASK(1)
+#define PRIVILEGE_TYPE_WRITE PRIVILEGE_TYPE_MASK(2)
+
+#define PRIVILEGE_TYPE_TEST_MASK(val, mask) (((val) & (mask)) != 0)
+
 typedef struct SDatabaseOptions {
   ENodeType   type;
   int32_t     buffer;
@@ -316,14 +324,6 @@ typedef struct SDropFunctionStmt {
   bool      ignoreNotExists;
 } SDropFunctionStmt;
 
-#define PRIVILEGE_TYPE_MASK(n) (1 << n)
-
-#define PRIVILEGE_TYPE_ALL   PRIVILEGE_TYPE_MASK(0)
-#define PRIVILEGE_TYPE_READ  PRIVILEGE_TYPE_MASK(1)
-#define PRIVILEGE_TYPE_WRITE PRIVILEGE_TYPE_MASK(2)
-
-#define PRIVILEGE_TYPE_TEST_MASK(val, mask) (((val) & (mask)) != 0)
-
 typedef struct SGrantStmt {
   ENodeType type;
   char      userName[TSDB_USER_LEN];
@@ -333,6 +333,30 @@ typedef struct SGrantStmt {
 
 typedef SGrantStmt SRevokeStmt;
 
+typedef struct SBalanceVgroupStmt {
+  ENodeType type;
+} SBalanceVgroupStmt;
+
+typedef struct SMergeVgroupStmt {
+  ENodeType type;
+  int32_t   vgId1;
+  int32_t   vgId2;
+} SMergeVgroupStmt;
+
+typedef struct SRedistributeVgroupStmt {
+  ENodeType  type;
+  int32_t    vgId;
+  int32_t    dnodeId1;
+  int32_t    dnodeId2;
+  int32_t    dnodeId3;
+  SNodeList* pDnodes;
+} SRedistributeVgroupStmt;
+
+typedef struct SSplitVgroupStmt {
+  ENodeType type;
+  int32_t   vgId;
+} SSplitVgroupStmt;
+
 #ifdef __cplusplus
 }
 #endif

include/libs/nodes/nodes.h

@@ -140,6 +140,7 @@ typedef enum ENodeType {
   QUERY_NODE_DROP_FUNCTION_STMT,
   QUERY_NODE_CREATE_STREAM_STMT,
   QUERY_NODE_DROP_STREAM_STMT,
+  QUERY_NODE_BALANCE_VGROUP_STMT,
   QUERY_NODE_MERGE_VGROUP_STMT,
   QUERY_NODE_REDISTRIBUTE_VGROUP_STMT,
   QUERY_NODE_SPLIT_VGROUP_STMT,

include/libs/stream/tstream.h

@@ -145,26 +145,6 @@ SStreamDataSubmit* streamSubmitRefClone(SStreamDataSubmit* pSubmit);
 #if 0
 int32_t streamDataBlockEncode(void** buf, const SStreamDataBlock* pOutput);
 void*   streamDataBlockDecode(const void* buf, SStreamDataBlock* pInput);
-
-static FORCE_INLINE int32_t streamEnqueue1(SStreamQueue* queue, SStreamQueueItem* pItem) {
-  int8_t inputStatus = atomic_load_8(&queue->enqueueStatus);
-  if (inputStatus == TASK_INPUT_STATUS__NORMAL) {
-    if (pItem->type == STREAM_INPUT__DATA_SUBMIT) {
-      SStreamDataSubmit* pSubmitClone = streamSubmitRefClone((SStreamDataSubmit*)pItem);
-      if (pSubmitClone == NULL) {
-        atomic_store_8(&queue->enqueueStatus, TASK_INPUT_STATUS__FAILED);
-        return -1;
-      }
-      taosWriteQitem(queue->queue, pSubmitClone);
-    } else if (pItem->type == STREAM_INPUT__DATA_BLOCK) {
-      taosWriteQitem(queue->queue, pItem);
-    } else if (pItem->type == STREAM_INPUT__CHECKPOINT) {
-      taosWriteQitem(queue->queue, pItem);
-    }
-    return 0;
-  }
-  return 0;
-}
 #endif
 
 typedef struct {

include/util/taoserror.h

@@ -653,6 +653,7 @@ int32_t* taosGetErrno();
 #define TSDB_CODE_PAR_VALUE_TOO_LONG            TAOS_DEF_ERROR_CODE(0, 0x2653)
 #define TSDB_CODE_PAR_INTERNAL_ERROR            TAOS_DEF_ERROR_CODE(0, 0x2654)
 #define TSDB_CODE_PAR_INVALID_DELETE_WHERE      TAOS_DEF_ERROR_CODE(0, 0x2655)
+#define TSDB_CODE_PAR_INVALID_REDISTRIBUTE_VG   TAOS_DEF_ERROR_CODE(0, 0x2656)
 
 //planner
 #define TSDB_CODE_PLAN_INTERNAL_ERROR           TAOS_DEF_ERROR_CODE(0, 0x2700)

source/client/src/tmq.c

@@ -195,7 +195,7 @@ typedef struct {
 
 tmq_conf_t* tmq_conf_new() {
   tmq_conf_t* conf = taosMemoryCalloc(1, sizeof(tmq_conf_t));
-  conf->withTbName = -1;
+  conf->withTbName = false;
   conf->autoCommit = true;
   conf->autoCommitInterval = 5000;
   conf->resetOffset = TMQ_CONF__RESET_OFFSET__EARLIEAST;
@@ -256,13 +256,10 @@ tmq_conf_res_t tmq_conf_set(tmq_conf_t* conf, const char* key, const char* value
 
   if (strcmp(key, "msg.with.table.name") == 0) {
     if (strcmp(value, "true") == 0) {
-      conf->withTbName = 1;
+      conf->withTbName = true;
       return TMQ_CONF_OK;
     } else if (strcmp(value, "false") == 0) {
-      conf->withTbName = 0;
-      return TMQ_CONF_OK;
-    } else if (strcmp(value, "none") == 0) {
-      conf->withTbName = -1;
+      conf->withTbName = false;
       return TMQ_CONF_OK;
     } else {
       return TMQ_CONF_INVALID;

source/common/src/tdatablock.c

@@ -18,6 +18,7 @@
 #include "tcompare.h"
 #include "tglobal.h"
 #include "tlog.h"
+#include "tname.h"
 
 int32_t taosGetFqdnPortFromEp(const char* ep, SEp* pEp) {
   pEp->port = 0;
@@ -99,7 +100,7 @@ void colDataTrim(SColumnInfoData* pColumnInfoData) {
   // TODO
 }
 
-int32_t getJsonValueLen(const char *data) {
+int32_t getJsonValueLen(const char* data) {
   int32_t dataLen = 0;
   if (*data == TSDB_DATA_TYPE_NULL) {
     dataLen = CHAR_BYTES;
@@ -109,7 +110,7 @@ int32_t getJsonValueLen(const char *data) {
     dataLen = DOUBLE_BYTES + CHAR_BYTES;
   } else if (*data == TSDB_DATA_TYPE_BOOL) {
     dataLen = CHAR_BYTES + CHAR_BYTES;
-  } else if (*data & TD_TAG_JSON) {   // json string
+  } else if (*data & TD_TAG_JSON) {  // json string
     dataLen = ((STag*)(data))->len;
   } else {
     ASSERT(0);
@@ -137,7 +138,7 @@ int32_t colDataAppend(SColumnInfoData* pColumnInfoData, uint32_t currentRow, con
     int32_t dataLen = 0;
     if (type == TSDB_DATA_TYPE_JSON) {
       dataLen = getJsonValueLen(pData);
-    }else {
+    } else {
       dataLen = varDataTLen(pData);
     }
 
@@ -1283,7 +1284,7 @@ static void doShiftBitmap(char* nullBitmap, size_t n, size_t total) {
   if (n % 8 == 0) {
     memmove(nullBitmap, nullBitmap + n / 8, newLen);
   } else {
-    int32_t tail = n % 8;
+    int32_t  tail = n % 8;
     int32_t  i = 0;
     uint8_t* p = (uint8_t*)nullBitmap;
 
@@ -1301,7 +1302,7 @@ static void doShiftBitmap(char* nullBitmap, size_t n, size_t total) {
       }
     } else if (n > 8) {
       int32_t gap = len - newLen;
-      while(i < newLen) {
+      while (i < newLen) {
         uint8_t v = p[i + gap];
         p[i] = (v << tail);
 
@@ -1316,7 +1317,6 @@ static void doShiftBitmap(char* nullBitmap, size_t n, size_t total) {
   }
 }
 
-
 static void colDataTrimFirstNRows(SColumnInfoData* pColInfoData, size_t n, size_t total) {
   if (IS_VAR_DATA_TYPE(pColInfoData->info.type)) {
     memmove(pColInfoData->varmeta.offset, &pColInfoData->varmeta.offset[n], (total - n) * sizeof(int32_t));
@@ -1544,7 +1544,8 @@ void blockDebugShowData(const SArray* dataBlocks, const char* flag) {
  *
  * TODO: colId should be set
  */
-int32_t buildSubmitReqFromDataBlock(SSubmitReq** pReq, const SArray* pDataBlocks, STSchema* pTSchema, int32_t vgId, tb_uid_t suid) {
+int32_t buildSubmitReqFromDataBlock(SSubmitReq** pReq, const SArray* pDataBlocks, STSchema* pTSchema, int32_t vgId,
+                                    tb_uid_t suid) {
   int32_t sz = taosArrayGetSize(pDataBlocks);
   int32_t bufSize = sizeof(SSubmitReq);
   for (int32_t i = 0; i < sz; ++i) {
@@ -1585,12 +1586,12 @@ int32_t buildSubmitReqFromDataBlock(SSubmitReq** pReq, const SArray* pDataBlocks
     int32_t dataLen = 0;
     for (int32_t j = 0; j < rows; ++j) {                     // iterate by row
       tdSRowResetBuf(&rb, POINTER_SHIFT(pDataBuf, msgLen));  // set row buf
-      bool isStartKey = false;
+      bool    isStartKey = false;
       int32_t offset = 0;
       for (int32_t k = 0; k < colNum; ++k) {  // iterate by column
         SColumnInfoData* pColInfoData = taosArrayGet(pDataBlock->pDataBlock, k);
-        STColumn*         pCol = &pTSchema->columns[k];
-        void* var = POINTER_SHIFT(pColInfoData->pData, j * pColInfoData->info.bytes);
+        STColumn*        pCol = &pTSchema->columns[k];
+        void*            var = POINTER_SHIFT(pColInfoData->pData, j * pColInfoData->info.bytes);
         switch (pColInfoData->info.type) {
           case TSDB_DATA_TYPE_TIMESTAMP:
             if (!isStartKey) {
@@ -1599,15 +1600,18 @@ int32_t buildSubmitReqFromDataBlock(SSubmitReq** pReq, const SArray* pDataBlocks
                                   offset, k);
 
             } else {
-              tdAppendColValToRow(&rb, PRIMARYKEY_TIMESTAMP_COL_ID + k, TSDB_DATA_TYPE_TIMESTAMP, TD_VTYPE_NORM, var, true, offset, k);
+              tdAppendColValToRow(&rb, PRIMARYKEY_TIMESTAMP_COL_ID + k, TSDB_DATA_TYPE_TIMESTAMP, TD_VTYPE_NORM, var,
+                                  true, offset, k);
             }
             break;
           case TSDB_DATA_TYPE_NCHAR: {
-            tdAppendColValToRow(&rb, PRIMARYKEY_TIMESTAMP_COL_ID + k, TSDB_DATA_TYPE_NCHAR, TD_VTYPE_NORM, var, true, offset, k);
+            tdAppendColValToRow(&rb, PRIMARYKEY_TIMESTAMP_COL_ID + k, TSDB_DATA_TYPE_NCHAR, TD_VTYPE_NORM, var, true,
+                                offset, k);
             break;
           }
           case TSDB_DATA_TYPE_VARCHAR: {  // TSDB_DATA_TYPE_BINARY
-            tdAppendColValToRow(&rb, PRIMARYKEY_TIMESTAMP_COL_ID + k, TSDB_DATA_TYPE_VARCHAR, TD_VTYPE_NORM, var, true, offset, k);
+            tdAppendColValToRow(&rb, PRIMARYKEY_TIMESTAMP_COL_ID + k, TSDB_DATA_TYPE_VARCHAR, TD_VTYPE_NORM, var, true,
+                                offset, k);
             break;
           }
           case TSDB_DATA_TYPE_VARBINARY:
@@ -1645,7 +1649,7 @@ int32_t buildSubmitReqFromDataBlock(SSubmitReq** pReq, const SArray* pDataBlocks
             }
             break;
         }
-        offset += TYPE_BYTES[pCol->type]; // sum/avg would convert to int64_t/uint64_t/double during aggregation
+        offset += TYPE_BYTES[pCol->type];  // sum/avg would convert to int64_t/uint64_t/double during aggregation
       }
       dataLen += TD_ROW_LEN(rb.pBuf);
 #ifdef TD_DEBUG_PRINT_ROW
@@ -1681,11 +1685,38 @@ int32_t buildSubmitReqFromDataBlock(SSubmitReq** pReq, const SArray* pDataBlocks
   return TSDB_CODE_SUCCESS;
 }
 
+char* buildCtbNameByGroupId(const char* stbName, uint64_t groupId) {
+  SArray* tags = taosArrayInit(0, sizeof(void*));
+  SSmlKv* pTag = taosMemoryCalloc(1, sizeof(SSmlKv));
+  pTag->key = "group_id";
+  pTag->keyLen = strlen(pTag->key);
+  pTag->type = TSDB_DATA_TYPE_UBIGINT;
+  pTag->u = groupId;
+  taosArrayPush(tags, &pTag);
+
+  void* cname = taosMemoryCalloc(1, TSDB_TABLE_NAME_LEN + 1);
+
+  RandTableName rname = {
+      .tags = tags,
+      .sTableName = stbName,
+      .sTableNameLen = strlen(stbName),
+      .childTableName = cname,
+  };
+
+  buildChildTableName(&rname);
+
+  taosMemoryFree(pTag);
+  taosArrayDestroy(tags);
+
+  ASSERT(rname.childTableName && rname.childTableName[0]);
+  return rname.childTableName;
+}
+
 SSubmitReq* tdBlockToSubmit(const SArray* pBlocks, const STSchema* pTSchema, bool createTb, int64_t suid,
                             const char* stbFullName, int32_t vgId) {
   SSubmitReq* ret = NULL;
   SArray*     tagArray = taosArrayInit(1, sizeof(STagVal));
-  if(!tagArray) {
+  if (!tagArray) {
     terrno = TSDB_CODE_OUT_OF_MEMORY;
     return NULL;
   }
@@ -1703,15 +1734,12 @@ SSubmitReq* tdBlockToSubmit(const SArray* pBlocks, const STSchema* pTSchema, boo
 
     if (createTb) {
       SVCreateTbReq createTbReq = {0};
-      char*         cname = taosMemoryCalloc(1, TSDB_TABLE_FNAME_LEN);
-      snprintf(cname, TSDB_TABLE_FNAME_LEN, "%s:%ld", stbFullName, pDataBlock->info.groupId);
+      char*         cname = buildCtbNameByGroupId(stbFullName, pDataBlock->info.groupId);
       createTbReq.name = cname;
       createTbReq.flags = 0;
       createTbReq.type = TSDB_CHILD_TABLE;
       createTbReq.ctb.suid = suid;
 
-
-
       STagVal tagVal = {.cid = 1,
                         .type = TSDB_DATA_TYPE_UBIGINT,
                         .pData = (uint8_t*)&pDataBlock->info.groupId,

source/common/src/tmsg.c

@@ -3258,6 +3258,89 @@ int32_t tDeserializeSKillTransReq(void *buf, int32_t bufLen, SKillTransReq *pReq
   return 0;
 }
 
+int32_t tSerializeSBalanceVgroupReq(void *buf, int32_t bufLen, SBalanceVgroupReq *pReq) {
+  SEncoder encoder = {0};
+  tEncoderInit(&encoder, buf, bufLen);
+
+  if (tStartEncode(&encoder) < 0) return -1;
+  if (tEncodeI32(&encoder, pReq->useless) < 0) return -1;
+  tEndEncode(&encoder);
+
+  int32_t tlen = encoder.pos;
+  tEncoderClear(&encoder);
+  return tlen;
+}
+
+int32_t tDeserializeSBalanceVgroupReq(void *buf, int32_t bufLen, SBalanceVgroupReq *pReq) {
+  SDecoder decoder = {0};
+  tDecoderInit(&decoder, buf, bufLen);
+
+  if (tStartDecode(&decoder) < 0) return -1;
+  if (tDecodeI32(&decoder, &pReq->useless) < 0) return -1;
+  tEndDecode(&decoder);
+
+  tDecoderClear(&decoder);
+  return 0;
+}
+
+int32_t tSerializeSMergeVgroupReq(void *buf, int32_t bufLen, SMergeVgroupReq *pReq) {
+  SEncoder encoder = {0};
+  tEncoderInit(&encoder, buf, bufLen);
+
+  if (tStartEncode(&encoder) < 0) return -1;
+  if (tEncodeI32(&encoder, pReq->vgId1) < 0) return -1;
+  if (tEncodeI32(&encoder, pReq->vgId2) < 0) return -1;
+  tEndEncode(&encoder);
+
+  int32_t tlen = encoder.pos;
+  tEncoderClear(&encoder);
+  return tlen;
+}
+
+int32_t tDeserializeSMergeVgroupReq(void *buf, int32_t bufLen, SMergeVgroupReq *pReq) {
+  SDecoder decoder = {0};
+  tDecoderInit(&decoder, buf, bufLen);
+
+  if (tStartDecode(&decoder) < 0) return -1;
+  if (tDecodeI32(&decoder, &pReq->vgId1) < 0) return -1;
+  if (tDecodeI32(&decoder, &pReq->vgId2) < 0) return -1;
+  tEndDecode(&decoder);
+
+  tDecoderClear(&decoder);
+  return 0;
+}
+
+int32_t tSerializeSRedistributeVgroupReq(void *buf, int32_t bufLen, SRedistributeVgroupReq *pReq) {
+  SEncoder encoder = {0};
+  tEncoderInit(&encoder, buf, bufLen);
+
+  if (tStartEncode(&encoder) < 0) return -1;
+  if (tEncodeI32(&encoder, pReq->vgId) < 0) return -1;
+  if (tEncodeI32(&encoder, pReq->dnodeId1) < 0) return -1;
+  if (tEncodeI32(&encoder, pReq->dnodeId2) < 0) return -1;
+  if (tEncodeI32(&encoder, pReq->dnodeId3) < 0) return -1;
+  tEndEncode(&encoder);
+
+  int32_t tlen = encoder.pos;
+  tEncoderClear(&encoder);
+  return tlen;
+}
+
+int32_t tDeserializeSRedistributeVgroupReq(void *buf, int32_t bufLen, SRedistributeVgroupReq *pReq) {
+  SDecoder decoder = {0};
+  tDecoderInit(&decoder, buf, bufLen);
+
+  if (tStartDecode(&decoder) < 0) return -1;
+  if (tDecodeI32(&decoder, &pReq->vgId) < 0) return -1;
+  if (tDecodeI32(&decoder, &pReq->dnodeId1) < 0) return -1;
+  if (tDecodeI32(&decoder, &pReq->dnodeId2) < 0) return -1;
+  if (tDecodeI32(&decoder, &pReq->dnodeId3) < 0) return -1;
+  tEndDecode(&decoder);
+
+  tDecoderClear(&decoder);
+  return 0;
+}
+
 int32_t tSerializeSDCreateMnodeReq(void *buf, int32_t bufLen, SDCreateMnodeReq *pReq) {
   SEncoder encoder = {0};
   tEncoderInit(&encoder, buf, bufLen);

source/dnode/vnode/CMakeLists.txt

@@ -36,12 +36,12 @@ target_sources(
 
     # tsdb
     "src/tsdb/tsdbCommit.c"
-    "src/tsdb/tsdbCommit2.c"
+    # "src/tsdb/tsdbCommit2.c"
     "src/tsdb/tsdbFile.c"
     "src/tsdb/tsdbFS.c"
     "src/tsdb/tsdbOpen.c"
     "src/tsdb/tsdbMemTable.c"
-    "src/tsdb/tsdbMemTable2.c"
+    # "src/tsdb/tsdbMemTable2.c"
     "src/tsdb/tsdbRead.c"
     "src/tsdb/tsdbReadImpl.c"
     "src/tsdb/tsdbWrite.c"

source/dnode/vnode/src/inc/tsdb.h

@@ -39,24 +39,46 @@ typedef struct SDelOp  SDelOp;
 
 static int tsdbKeyCmprFn(const void *p1, const void *p2);
 
+// tsdbMemTable ==============================================================================================
+typedef struct STbData     STbData;
+typedef struct SMemTable   SMemTable;
+typedef struct STbDataIter STbDataIter;
+typedef struct SMergeInfo  SMergeInfo;
+typedef struct STable      STable;
+
+// SMemTable
+int32_t tsdbMemTableCreate(STsdb *pTsdb, SMemTable **ppMemTable);
+void    tsdbMemTableDestroy(SMemTable *pMemTable);
+void    tsdbGetTbDataFromMemTable(SMemTable *pMemTable, tb_uid_t suid, tb_uid_t uid, STbData **ppTbData);
+
+// STbDataIter
+int32_t tsdbTbDataIterCreate(STbData *pTbData, TSDBKEY *pFrom, int8_t backward, STbDataIter **ppIter);
+void   *tsdbTbDataIterDestroy(STbDataIter *pIter);
+void    tsdbTbDataIterOpen(STbData *pTbData, TSDBKEY *pFrom, int8_t backward, STbDataIter *pIter);
+bool    tsdbTbDataIterNext(STbDataIter *pIter);
+bool    tsdbTbDataIterGet(STbDataIter *pIter, TSDBROW *pRow);
+
+int tsdbLoadDataFromCache(STsdb *pTsdb, STable *pTable, STbDataIter *pIter, TSKEY maxKey, int maxRowsToRead,
+                          SDataCols *pCols, TKEY *filterKeys, int nFilterKeys, bool keepDup, SMergeInfo *pMergeInfo);
+
 // tsdbMemTable2.c ==============================================================================================
-typedef struct SMemTable    SMemTable;
-typedef struct SMemData     SMemData;
-typedef struct SMemDataIter SMemDataIter;
+// typedef struct SMemTable2   SMemTable2;
+// typedef struct SMemData     SMemData;
+// typedef struct SMemDataIter SMemDataIter;
 
-int32_t tsdbMemTableCreate2(STsdb *pTsdb, SMemTable **ppMemTable);
-void    tsdbMemTableDestroy2(SMemTable *pMemTable);
-int32_t tsdbInsertTableData2(STsdb *pTsdb, int64_t version, SVSubmitBlk *pSubmitBlk);
-int32_t tsdbDeleteTableData2(STsdb *pTsdb, int64_t version, tb_uid_t suid, tb_uid_t uid, TSKEY sKey, TSKEY eKey);
+// int32_t tsdbMemTableCreate2(STsdb *pTsdb, SMemTable2 **ppMemTable);
+// void    tsdbMemTableDestroy2(SMemTable2 *pMemTable);
+// int32_t tsdbInsertTableData2(STsdb *pTsdb, int64_t version, SVSubmitBlk *pSubmitBlk);
+// int32_t tsdbDeleteTableData2(STsdb *pTsdb, int64_t version, tb_uid_t suid, tb_uid_t uid, TSKEY sKey, TSKEY eKey);
 
-/* SMemDataIter */
-void tsdbMemDataIterOpen(SMemData *pMemData, TSDBKEY *pKey, int8_t backward, SMemDataIter *pIter);
-bool tsdbMemDataIterNext(SMemDataIter *pIter);
-void tsdbMemDataIterGet(SMemDataIter *pIter, TSDBROW **ppRow);
+// /* SMemDataIter */
+// void tsdbMemDataIterOpen(SMemData *pMemData, TSDBKEY *pKey, int8_t backward, SMemDataIter *pIter);
+// bool tsdbMemDataIterNext(SMemDataIter *pIter);
+// void tsdbMemDataIterGet(SMemDataIter *pIter, TSDBROW **ppRow);
 
-// tsdbCommit2.c ==============================================================================================
-int32_t tsdbBegin2(STsdb *pTsdb);
-int32_t tsdbCommit2(STsdb *pTsdb);
+// // tsdbCommit2.c ==============================================================================================
+// int32_t tsdbBegin2(STsdb *pTsdb);
+// int32_t tsdbCommit2(STsdb *pTsdb);
 
 // tsdbFile.c ==============================================================================================
 typedef int32_t          TSDB_FILE_T;
@@ -124,17 +146,6 @@ int        tsdbRLockFS(STsdbFS *pFs);
 int        tsdbWLockFS(STsdbFS *pFs);
 int        tsdbUnLockFS(STsdbFS *pFs);
 
-// tsdbMemTable ================
-typedef struct STbData       STbData;
-typedef struct STsdbMemTable STsdbMemTable;
-typedef struct SMergeInfo    SMergeInfo;
-typedef struct STable        STable;
-
-int  tsdbMemTableCreate(STsdb *pTsdb, STsdbMemTable **ppMemTable);
-void tsdbMemTableDestroy(STsdbMemTable *pMemTable);
-int  tsdbLoadDataFromCache(STsdb *pTsdb, STable *pTable, SSkipListIterator *pIter, TSKEY maxKey, int maxRowsToRead,
-                           SDataCols *pCols, TKEY *filterKeys, int nFilterKeys, bool keepDup, SMergeInfo *pMergeInfo);
-
 // structs
 typedef struct {
   int   minFid;
@@ -145,30 +156,27 @@ typedef struct {
 
 #define TSDB_DATA_DIR_LEN 6  // adapt accordingly
 struct STsdb {
-  char          *path;
-  SVnode        *pVnode;
-  TdThreadMutex  mutex;
-  char           dir[TSDB_DATA_DIR_LEN];
-  bool           repoLocked;
-  STsdbKeepCfg   keepCfg;
-  STsdbMemTable *mem;
-  STsdbMemTable *imem;
-  SRtn           rtn;
-  STsdbFS       *fs;
+  char         *path;
+  SVnode       *pVnode;
+  TdThreadMutex mutex;
+  char          dir[TSDB_DATA_DIR_LEN];
+  bool          repoLocked;
+  STsdbKeepCfg  keepCfg;
+  SMemTable    *mem;
+  SMemTable    *imem;
+  SRtn          rtn;
+  STsdbFS      *fs;
 };
 
 #if 1  // ======================================
 
 struct STable {
-  uint64_t  tid;
+  uint64_t  suid;
   uint64_t  uid;
   STSchema *pSchema;       // latest schema
   STSchema *pCacheSchema;  // cached cache
 };
 
-#define TABLE_TID(t) (t)->tid
-#define TABLE_UID(t) (t)->uid
-
 // int tsdbPrepareCommit(STsdb *pTsdb);
 typedef enum {
   TSDB_FILE_HEAD = 0,  // .head
@@ -206,27 +214,44 @@ struct SDFileSet {
   SDFile   files[TSDB_FILE_MAX];
 };
 
+struct TSDBKEY {
+  int64_t version;
+  TSKEY   ts;
+};
+
+typedef struct SMemSkipListNode SMemSkipListNode;
+struct SMemSkipListNode {
+  int8_t            level;
+  SMemSkipListNode *forwards[0];
+};
+typedef struct SMemSkipList {
+  uint32_t          seed;
+  int64_t           size;
+  int8_t            maxLevel;
+  int8_t            level;
+  SMemSkipListNode *pHead;
+  SMemSkipListNode *pTail;
+} SMemSkipList;
+
 struct STbData {
-  tb_uid_t   uid;
-  TSKEY      keyMin;
-  TSKEY      keyMax;
-  int64_t    minVer;
-  int64_t    maxVer;
-  int64_t    nrows;
-  SSkipList *pData;
+  tb_uid_t     suid;
+  tb_uid_t     uid;
+  TSDBKEY      minKey;
+  TSDBKEY      maxKey;
+  SDelOp      *pHead;
+  SDelOp      *pTail;
+  SMemSkipList sl;
 };
 
-struct STsdbMemTable {
-  SVBufPool *pPool;
-  T_REF_DECLARE()
-  SRWLatch   latch;
-  TSKEY      keyMin;
-  TSKEY      keyMax;
-  int64_t    minVer;
-  int64_t    maxVer;
-  int64_t    nRow;
-  SSkipList *pSlIdx;  // SSkiplist<STbData>
-  SHashObj  *pHashIdx;
+struct SMemTable {
+  SRWLatch latch;
+  STsdb   *pTsdb;
+  int32_t  nRef;
+  TSDBKEY  minKey;
+  TSDBKEY  maxKey;
+  int64_t  nRow;
+  int64_t  nDelOp;
+  SArray  *aTbData;  // SArray<STbData>
 };
 
 struct STsdbFSMeta {
@@ -237,9 +262,11 @@ struct STsdbFSMeta {
 
 // ==================
 typedef struct {
-  STsdbFSMeta meta;  // FS meta
-  SArray     *df;    // data file array
-  SArray     *sf;    // sma data file array    v2f1900.index_name_1
+  STsdbFSMeta meta;       // FS meta
+  SDFile      cacheFile;  // cache file
+  SDFile      tombstone;  // tomestome file
+  SArray     *df;         // data file array
+  SArray     *sf;         // sma data file array    v2f1900.index_name_1
 } SFSStatus;
 
 struct STsdbFS {
@@ -292,16 +319,24 @@ static void  *taosTZfree(void *ptr);
 static size_t taosTSizeof(void *ptr);
 static void   taosTMemset(void *ptr, int c);
 
-static FORCE_INLINE STSRow *tsdbNextIterRow(SSkipListIterator *pIter) {
+struct TSDBROW {
+  int64_t version;
+  STSRow *pTSRow;
+};
+
+static FORCE_INLINE STSRow *tsdbNextIterRow(STbDataIter *pIter) {
+  TSDBROW row;
+
   if (pIter == NULL) return NULL;
 
-  SSkipListNode *node = tSkipListIterGet(pIter);
-  if (node == NULL) return NULL;
+  if (tsdbTbDataIterGet(pIter, &row)) {
+    return row.pTSRow;
+  }
 
-  return (STSRow *)SL_GET_NODE_DATA(node);
+  return NULL;
 }
 
-static FORCE_INLINE TSKEY tsdbNextIterKey(SSkipListIterator *pIter) {
+static FORCE_INLINE TSKEY tsdbNextIterKey(STbDataIter *pIter) {
   STSRow *row = tsdbNextIterRow(pIter);
   if (row == NULL) return TSDB_DATA_TIMESTAMP_NULL;
 
@@ -311,11 +346,6 @@ static FORCE_INLINE TSKEY tsdbNextIterKey(SSkipListIterator *pIter) {
 // tsdbReadImpl
 typedef struct SReadH SReadH;
 
-struct TSDBKEY {
-  int64_t version;
-  TSKEY   ts;
-};
-
 typedef struct {
   uint64_t suid;
   uint64_t uid;
@@ -354,7 +384,7 @@ typedef struct {
 
 typedef struct {
   int32_t  delimiter;  // For recovery usage
-  int32_t  tid;
+  uint64_t suid;
   uint64_t uid;
   SBlock   blocks[];
 } SBlockInfo;
@@ -650,11 +680,6 @@ struct SFSIter {
 #define TSDB_FS_ITER_FORWARD  TSDB_ORDER_ASC
 #define TSDB_FS_ITER_BACKWARD TSDB_ORDER_DESC
 
-struct TSDBROW {
-  int64_t version;
-  STSRow2 tsRow;
-};
-
 struct TABLEID {
   tb_uid_t suid;
   tb_uid_t uid;
@@ -675,7 +700,7 @@ typedef struct {
   TSKEY    eKey;
 } SDelInfo;
 
-struct SMemTable {
+struct SMemTable2 {
   STsdb  *pTsdb;
   int32_t nRef;
   TSDBKEY minKey;
@@ -705,16 +730,6 @@ static FORCE_INLINE int tsdbKeyCmprFn(const void *p1, const void *p2) {
   return 0;
 }
 
-typedef struct SMemSkipListNode SMemSkipListNode;
-typedef struct SMemSkipList {
-  uint32_t          seed;
-  int32_t           size;
-  int8_t            maxLevel;
-  int8_t            level;
-  SMemSkipListNode *pHead;
-  SMemSkipListNode *pTail;
-} SMemSkipList;
-
 struct SMemData {
   tb_uid_t     suid;
   tb_uid_t     uid;
@@ -726,13 +741,19 @@ struct SMemData {
 };
 
 struct SMemDataIter {
-  SMemData         *pMemData;
+  STbData          *pMemData;
   int8_t            backward;
   TSDBROW          *pRow;
   SMemSkipListNode *pNode;  // current node
   TSDBROW           row;
 };
 
+struct STbDataIter {
+  STbData          *pTbData;
+  int8_t            backward;
+  SMemSkipListNode *pNode;
+};
+
 #endif
 
 #ifdef __cplusplus

source/dnode/vnode/src/inc/vnodeInt.h

@@ -116,7 +116,9 @@ int          tsdbBegin(STsdb* pTsdb);
 int32_t      tsdbCommit(STsdb* pTsdb);
 int          tsdbScanAndConvertSubmitMsg(STsdb* pTsdb, SSubmitReq* pMsg);
 int          tsdbInsertData(STsdb* pTsdb, int64_t version, SSubmitReq* pMsg, SSubmitRsp* pRsp);
-int          tsdbInsertTableData(STsdb* pTsdb, SSubmitMsgIter* pMsgIter, SSubmitBlk* pBlock, SSubmitBlkRsp* pRsp);
+int32_t      tsdbInsertTableData(STsdb* pTsdb, int64_t version, SSubmitMsgIter* pMsgIter, SSubmitBlk* pBlock,
+                                 SSubmitBlkRsp* pRsp);
+int32_t      tsdbDeleteTableData(STsdb* pTsdb, int64_t version, tb_uid_t suid, tb_uid_t uid, TSKEY sKey, TSKEY eKey);
 tsdbReaderT* tsdbQueryTables(SVnode* pVnode, SQueryTableDataCond* pCond, STableListInfo* tableList, uint64_t qId,
                              uint64_t taskId);
 tsdbReaderT  tsdbQueryCacheLastT(STsdb* tsdb, SQueryTableDataCond* pCond, STableListInfo* tableList, uint64_t qId,

source/dnode/vnode/src/meta/metaQuery.c

@@ -31,7 +31,7 @@ void metaReaderClear(SMetaReader *pReader) {
 }
 
 int metaGetTableEntryByVersion(SMetaReader *pReader, int64_t version, tb_uid_t uid) {
-  SMeta *  pMeta = pReader->pMeta;
+  SMeta   *pMeta = pReader->pMeta;
   STbDbKey tbDbKey = {.version = version, .uid = uid};
 
   // query table.db
@@ -54,7 +54,7 @@ _err:
 }
 
 int metaGetTableEntryByUid(SMetaReader *pReader, tb_uid_t uid) {
-  SMeta * pMeta = pReader->pMeta;
+  SMeta  *pMeta = pReader->pMeta;
   int64_t version;
 
   // query uid.idx
@@ -68,7 +68,7 @@ int metaGetTableEntryByUid(SMetaReader *pReader, tb_uid_t uid) {
 }
 
 int metaGetTableEntryByName(SMetaReader *pReader, const char *name) {
-  SMeta *  pMeta = pReader->pMeta;
+  SMeta   *pMeta = pReader->pMeta;
   tb_uid_t uid;
 
   // query name.idx
@@ -82,7 +82,7 @@ int metaGetTableEntryByName(SMetaReader *pReader, const char *name) {
 }
 
 tb_uid_t metaGetTableEntryUidByName(SMeta *pMeta, const char *name) {
-  void *   pData = NULL;
+  void    *pData = NULL;
   int      nData = 0;
   tb_uid_t uid = 0;
 
@@ -134,7 +134,7 @@ void metaCloseTbCursor(SMTbCursor *pTbCur) {
 
 int metaTbCursorNext(SMTbCursor *pTbCur) {
   int    ret;
-  void * pBuf;
+  void  *pBuf;
   STbCfg tbCfg;
 
   for (;;) {
@@ -155,7 +155,7 @@ int metaTbCursorNext(SMTbCursor *pTbCur) {
 }
 
 SSchemaWrapper *metaGetTableSchema(SMeta *pMeta, tb_uid_t uid, int32_t sver, bool isinline) {
-  void *          pData = NULL;
+  void           *pData = NULL;
   int             nData = 0;
   int64_t         version;
   SSchemaWrapper  schema = {0};
@@ -163,37 +163,47 @@ SSchemaWrapper *metaGetTableSchema(SMeta *pMeta, tb_uid_t uid, int32_t sver, boo
   SDecoder        dc = {0};
 
   metaRLock(pMeta);
-  if (sver < 0) {
-    if (tdbTbGet(pMeta->pUidIdx, &uid, sizeof(uid), &pData, &nData) < 0) {
-      goto _err;
-    }
-
-    version = *(int64_t *)pData;
+_query:
+  if (tdbTbGet(pMeta->pUidIdx, &uid, sizeof(uid), &pData, &nData) < 0) {
+    goto _err;
+  }
 
-    tdbTbGet(pMeta->pTbDb, &(STbDbKey){.uid = uid, .version = version}, sizeof(STbDbKey), &pData, &nData);
+  version = *(int64_t *)pData;
 
-    SMetaEntry me = {0};
-    tDecoderInit(&dc, pData, nData);
-    metaDecodeEntry(&dc, &me);
-    if (me.type == TSDB_SUPER_TABLE) {
+  tdbTbGet(pMeta->pTbDb, &(STbDbKey){.uid = uid, .version = version}, sizeof(STbDbKey), &pData, &nData);
+  SMetaEntry me = {0};
+  tDecoderInit(&dc, pData, nData);
+  metaDecodeEntry(&dc, &me);
+  if (me.type == TSDB_SUPER_TABLE) {
+    if (sver == -1 || sver == me.stbEntry.schemaRow.version) {
       pSchema = tCloneSSchemaWrapper(&me.stbEntry.schemaRow);
-    } else if (me.type == TSDB_NORMAL_TABLE) {
-      pSchema = tCloneSSchemaWrapper(&me.ntbEntry.schemaRow);
-    } else {
-      ASSERT(0);
+      tDecoderClear(&dc);
+      goto _exit;
     }
+  } else if (me.type == TSDB_CHILD_TABLE) {
+    uid = me.ctbEntry.suid;
     tDecoderClear(&dc);
+    goto _query;
   } else {
-    if (tdbTbGet(pMeta->pSkmDb, &(SSkmDbKey){.uid = uid, .sver = sver}, sizeof(SSkmDbKey), &pData, &nData) < 0) {
-      goto _err;
+    if (sver == -1 || sver == me.ntbEntry.schemaRow.version) {
+      pSchema = tCloneSSchemaWrapper(&me.ntbEntry.schemaRow);
+      tDecoderClear(&dc);
+      goto _exit;
     }
+  }
+  tDecoderClear(&dc);
 
-    tDecoderInit(&dc, pData, nData);
-    tDecodeSSchemaWrapper(&dc, &schema);
-    pSchema = tCloneSSchemaWrapper(&schema);
-    tDecoderClear(&dc);
+  // query from skm db
+  if (tdbTbGet(pMeta->pSkmDb, &(SSkmDbKey){.uid = uid, .sver = sver}, sizeof(SSkmDbKey), &pData, &nData) < 0) {
+    goto _err;
   }
 
+  tDecoderInit(&dc, pData, nData);
+  tDecodeSSchemaWrapper(&dc, &schema);
+  pSchema = tCloneSSchemaWrapper(&schema);
+  tDecoderClear(&dc);
+
+_exit:
   metaULock(pMeta);
   tdbFree(pData);
   return pSchema;
@@ -205,11 +215,11 @@ _err:
 }
 
 struct SMCtbCursor {
-  SMeta *  pMeta;
-  TBC *    pCur;
+  SMeta   *pMeta;
+  TBC     *pCur;
   tb_uid_t suid;
-  void *   pKey;
-  void *   pVal;
+  void    *pKey;
+  void    *pVal;
   int      kLen;
   int      vLen;
 };
@@ -279,25 +289,13 @@ tb_uid_t metaCtbCursorNext(SMCtbCursor *pCtbCur) {
 }
 
 STSchema *metaGetTbTSchema(SMeta *pMeta, tb_uid_t uid, int32_t sver) {
-  tb_uid_t        quid;
-  SMetaReader     mr = {0};
-  STSchema *      pTSchema = NULL;
+  // SMetaReader     mr = {0};
+  STSchema       *pTSchema = NULL;
   SSchemaWrapper *pSW = NULL;
   STSchemaBuilder sb = {0};
-  SSchema *       pSchema;
-
-  metaReaderInit(&mr, pMeta, 0);
-  metaGetTableEntryByUid(&mr, uid);
-
-  if (mr.me.type == TSDB_CHILD_TABLE) {
-    quid = mr.me.ctbEntry.suid;
-  } else {
-    quid = uid;
-  }
-
-  metaReaderClear(&mr);
+  SSchema        *pSchema;
 
-  pSW = metaGetTableSchema(pMeta, quid, sver, 0);
+  pSW = metaGetTableSchema(pMeta, uid, sver, 0);
   if (!pSW) return NULL;
 
   tdInitTSchemaBuilder(&sb, pSW->version);
@@ -321,11 +319,11 @@ int metaGetTbNum(SMeta *pMeta) {
 }
 
 typedef struct {
-  SMeta *  pMeta;
-  TBC *    pCur;
+  SMeta   *pMeta;
+  TBC     *pCur;
   tb_uid_t uid;
-  void *   pKey;
-  void *   pVal;
+  void    *pKey;
+  void    *pVal;
   int      kLen;
   int      vLen;
 } SMSmaCursor;
@@ -397,7 +395,7 @@ tb_uid_t metaSmaCursorNext(SMSmaCursor *pSmaCur) {
 
 STSmaWrapper *metaGetSmaInfoByTable(SMeta *pMeta, tb_uid_t uid, bool deepCopy) {
   STSmaWrapper *pSW = NULL;
-  SArray *      pSmaIds = NULL;
+  SArray       *pSmaIds = NULL;
 
   if (!(pSmaIds = metaGetSmaIdsByTable(pMeta, uid))) {
     return NULL;
@@ -421,7 +419,7 @@ STSmaWrapper *metaGetSmaInfoByTable(SMeta *pMeta, tb_uid_t uid, bool deepCopy) {
   metaReaderInit(&mr, pMeta, 0);
   int64_t smaId;
   int     smaIdx = 0;
-  STSma * pTSma = NULL;
+  STSma  *pTSma = NULL;
   for (int i = 0; i < pSW->number; ++i) {
     smaId = *(tb_uid_t *)taosArrayGet(pSmaIds, i);
     if (metaGetTableEntryByUid(&mr, smaId) < 0) {
@@ -469,7 +467,7 @@ _err:
 }
 
 STSma *metaGetSmaInfoByIndex(SMeta *pMeta, int64_t indexUid) {
-  STSma *     pTSma = NULL;
+  STSma      *pTSma = NULL;
   SMetaReader mr = {0};
   metaReaderInit(&mr, pMeta, 0);
   if (metaGetTableEntryByUid(&mr, indexUid) < 0) {
@@ -491,7 +489,7 @@ STSma *metaGetSmaInfoByIndex(SMeta *pMeta, int64_t indexUid) {
 }
 
 SArray *metaGetSmaIdsByTable(SMeta *pMeta, tb_uid_t uid) {
-  SArray *    pUids = NULL;
+  SArray     *pUids = NULL;
   SSmaIdxKey *pSmaIdxKey = NULL;
 
   SMSmaCursor *pCur = metaOpenSmaCursor(pMeta, uid);
@@ -529,7 +527,7 @@ SArray *metaGetSmaIdsByTable(SMeta *pMeta, tb_uid_t uid) {
 }
 
 SArray *metaGetSmaTbUids(SMeta *pMeta) {
-  SArray *    pUids = NULL;
+  SArray     *pUids = NULL;
   SSmaIdxKey *pSmaIdxKey = NULL;
   tb_uid_t    lastUid = 0;
 
@@ -591,13 +589,13 @@ const void *metaGetTableTagVal(SMetaEntry *pEntry, int16_t type, STagVal *val) {
 }
 
 typedef struct {
-  SMeta *  pMeta;
-  TBC *    pCur;
+  SMeta   *pMeta;
+  TBC     *pCur;
   tb_uid_t suid;
   int16_t  cid;
   int16_t  type;
-  void *   pKey;
-  void *   pVal;
+  void    *pKey;
+  void    *pVal;
   int32_t  kLen;
   int32_t  vLen;
 } SIdxCursor;
@@ -621,7 +619,7 @@ int32_t metaFilteTableIds(SMeta *pMeta, SMetaFltParam *param, SArray *pUids) {
   int32_t     nKey = 0;
 
   int32_t nTagData = 0;
-  void *  tagData = NULL;
+  void   *tagData = NULL;
 
   if (IS_VAR_DATA_TYPE(param->type)) {
     tagData = varDataVal(param->val);
@@ -640,7 +638,7 @@ int32_t metaFilteTableIds(SMeta *pMeta, SMetaFltParam *param, SArray *pUids) {
     goto END;
   }
 
-  void *  entryKey = NULL, *entryVal = NULL;
+  void   *entryKey = NULL, *entryVal = NULL;
   int32_t nEntryKey, nEntryVal;
   while (1) {
     valid = tdbTbcGet(pCursor->pCur, (const void **)&entryKey, &nEntryKey, (const void **)&entryVal, &nEntryVal);

source/dnode/vnode/src/tq/tqRead.c

@@ -141,10 +141,10 @@ int32_t tqRetrieveDataBlock(SArray** ppCols, STqReadHandle* pHandle, uint64_t* p
     }
 
     // this interface use suid instead of uid
-    pHandle->pSchemaWrapper = metaGetTableSchema(pHandle->pVnodeMeta, pHandle->msgIter.suid, sversion, true);
+    pHandle->pSchemaWrapper = metaGetTableSchema(pHandle->pVnodeMeta, pHandle->msgIter.uid, sversion, true);
     if (pHandle->pSchemaWrapper == NULL) {
       tqWarn("cannot found schema wrapper for table: suid: %ld, version %d, possibly dropped table",
-             pHandle->msgIter.suid, pHandle->cachedSchemaVer);
+             pHandle->msgIter.uid, pHandle->cachedSchemaVer);
       /*ASSERT(0);*/
       terrno = TSDB_CODE_TQ_TABLE_SCHEMA_NOT_FOUND;
       return -1;

source/dnode/vnode/src/tsdb/tsdbCommit.c

@@ -18,8 +18,8 @@
 #define TSDB_MAX_SUBBLOCKS 8
 
 typedef struct {
-  STable            *pTable;
-  SSkipListIterator *pIter;
+  STable      *pTable;
+  STbDataIter *pIter;
 } SCommitIter;
 
 typedef struct {
@@ -58,26 +58,26 @@ typedef struct {
 #define TSDB_COMMIT_DEFAULT_ROWS(ch) TSDB_DEFAULT_BLOCK_ROWS(TSDB_COMMIT_REPO(ch)->pVnode->config.tsdbCfg.maxRows)
 #define TSDB_COMMIT_TXN_VERSION(ch)  FS_TXN_VERSION(REPO_FS(TSDB_COMMIT_REPO(ch)))
 
-static void tsdbStartCommit(STsdb *pRepo);
-static void tsdbEndCommit(STsdb *pTsdb, int eno);
-static int  tsdbInitCommitH(SCommitH *pCommith, STsdb *pRepo);
-static void tsdbSeekCommitIter(SCommitH *pCommith, TSKEY key);
-static int  tsdbNextCommitFid(SCommitH *pCommith);
-static void tsdbDestroyCommitH(SCommitH *pCommith);
-static int  tsdbCreateCommitIters(SCommitH *pCommith);
-static void tsdbDestroyCommitIters(SCommitH *pCommith);
-static int  tsdbCommitToFile(SCommitH *pCommith, SDFileSet *pSet, int fid);
-static void tsdbResetCommitFile(SCommitH *pCommith);
-static int  tsdbSetAndOpenCommitFile(SCommitH *pCommith, SDFileSet *pSet, int fid);
-static int  tsdbCommitToTable(SCommitH *pCommith, int tid);
-static bool tsdbCommitIsSameFile(SCommitH *pCommith, int bidx);
-static int  tsdbMoveBlkIdx(SCommitH *pCommith, SBlockIdx *pIdx);
-static int  tsdbSetCommitTable(SCommitH *pCommith, STable *pTable);
-static int  tsdbComparKeyBlock(const void *arg1, const void *arg2);
-static int  tsdbWriteBlockInfo(SCommitH *pCommih);
-static int  tsdbCommitMemData(SCommitH *pCommith, SCommitIter *pIter, TSKEY keyLimit, bool toData);
-static int  tsdbMergeMemData(SCommitH *pCommith, SCommitIter *pIter, int bidx);
-static int  tsdbMoveBlock(SCommitH *pCommith, int bidx);
+static void    tsdbStartCommit(STsdb *pRepo);
+static void    tsdbEndCommit(STsdb *pTsdb, int eno);
+static int     tsdbInitCommitH(SCommitH *pCommith, STsdb *pRepo);
+static void    tsdbSeekCommitIter(SCommitH *pCommith, TSKEY key);
+static int     tsdbNextCommitFid(SCommitH *pCommith);
+static void    tsdbDestroyCommitH(SCommitH *pCommith);
+static int32_t tsdbCreateCommitIters(SCommitH *pCommith);
+static void    tsdbDestroyCommitIters(SCommitH *pCommith);
+static int     tsdbCommitToFile(SCommitH *pCommith, SDFileSet *pSet, int fid);
+static void    tsdbResetCommitFile(SCommitH *pCommith);
+static int     tsdbSetAndOpenCommitFile(SCommitH *pCommith, SDFileSet *pSet, int fid);
+static int     tsdbCommitToTable(SCommitH *pCommith, int tid);
+static bool    tsdbCommitIsSameFile(SCommitH *pCommith, int bidx);
+static int     tsdbMoveBlkIdx(SCommitH *pCommith, SBlockIdx *pIdx);
+static int     tsdbSetCommitTable(SCommitH *pCommith, STable *pTable);
+static int     tsdbComparKeyBlock(const void *arg1, const void *arg2);
+static int     tsdbWriteBlockInfo(SCommitH *pCommih);
+static int     tsdbCommitMemData(SCommitH *pCommith, SCommitIter *pIter, TSKEY keyLimit, bool toData);
+static int     tsdbMergeMemData(SCommitH *pCommith, SCommitIter *pIter, int bidx);
+static int     tsdbMoveBlock(SCommitH *pCommith, int bidx);
 static int  tsdbCommitAddBlock(SCommitH *pCommith, const SBlock *pSupBlock, const SBlock *pSubBlocks, int nSubBlocks);
 static int  tsdbMergeBlockData(SCommitH *pCommith, SCommitIter *pIter, SDataCols *pDataCols, TSKEY keyLimit,
                                bool isLastOneBlock);
@@ -92,7 +92,7 @@ static int  tsdbApplyRtnOnFSet(STsdb *pRepo, SDFileSet *pSet, SRtn *pRtn);
 int tsdbBegin(STsdb *pTsdb) {
   if (!pTsdb) return 0;
 
-  STsdbMemTable *pMem;
+  SMemTable *pMem;
 
   if (tsdbMemTableCreate(pTsdb, &pTsdb->mem) < 0) {
     return -1;
@@ -244,7 +244,7 @@ void tsdbGetRtnSnap(STsdb *pRepo, SRtn *pRtn) {
 }
 
 static void tsdbStartCommit(STsdb *pRepo) {
-  STsdbMemTable *pMem = pRepo->imem;
+  SMemTable *pMem = pRepo->imem;
 
   tsdbInfo("vgId:%d, start to commit", REPO_ID(pRepo));
 
@@ -400,7 +400,7 @@ static int tsdbCommitToFile(SCommitH *pCommith, SDFileSet *pSet, int fid) {
       break;
     }
 
-    if (pIter && pIter->pTable && (!pIdx || (pIter->pTable->uid <= pIdx->uid))) {
+    if (pIter && pIter->pTable && (!pIdx || (pIter->pTable->suid <= pIdx->suid || pIter->pTable->uid <= pIdx->uid))) {
       if (tsdbCommitToTable(pCommith, mIter) < 0) {
         tsdbCloseCommitFile(pCommith, true);
         // revert the file change
@@ -453,57 +453,48 @@ static int tsdbCommitToFile(SCommitH *pCommith, SDFileSet *pSet, int fid) {
   return 0;
 }
 
-static int tsdbCreateCommitIters(SCommitH *pCommith) {
-  STsdb             *pRepo = TSDB_COMMIT_REPO(pCommith);
-  STsdbMemTable     *pMem = pRepo->imem;
-  SSkipListIterator *pSlIter;
-  SCommitIter       *pCommitIter;
-  SSkipListNode     *pNode;
-  STbData           *pTbData;
-  STSchema          *pTSchema = NULL;
+static int32_t tsdbCreateCommitIters(SCommitH *pCommith) {
+  int32_t      code = 0;
+  STsdb       *pRepo = TSDB_COMMIT_REPO(pCommith);
+  SMemTable   *pMem = pRepo->imem;
+  STbData     *pTbData;
+  SCommitIter *pCommitIter;
+  STSchema    *pTSchema = NULL;
 
-  pCommith->niters = SL_SIZE(pMem->pSlIdx);
+  pCommith->niters = taosArrayGetSize(pMem->aTbData);
   pCommith->iters = (SCommitIter *)taosMemoryCalloc(pCommith->niters, sizeof(SCommitIter));
   if (pCommith->iters == NULL) {
-    terrno = TSDB_CODE_TDB_OUT_OF_MEMORY;
-    return -1;
+    code = TSDB_CODE_OUT_OF_MEMORY;
+    goto _err;
   }
 
-  // Loop to create iters for each skiplist
-  pSlIter = tSkipListCreateIter(pMem->pSlIdx);
-  if (pSlIter == NULL) {
-    terrno = TSDB_CODE_TDB_OUT_OF_MEMORY;
-    return -1;
-  }
-  for (int i = 0; i < pCommith->niters; i++) {
-    tSkipListIterNext(pSlIter);
-    pNode = tSkipListIterGet(pSlIter);
-    pTbData = (STbData *)pNode->pData;
+  for (int32_t iIter = 0; iIter < pCommith->niters; iIter++) {
+    pTbData = (STbData *)taosArrayGetP(pMem->aTbData, iIter);
+    pCommitIter = &pCommith->iters[iIter];
 
-    pCommitIter = pCommith->iters + i;
     pTSchema = metaGetTbTSchema(REPO_META(pRepo), pTbData->uid, -1);
-
     if (pTSchema) {
-      pCommitIter->pIter = tSkipListCreateIter(pTbData->pData);
-      tSkipListIterNext(pCommitIter->pIter);
+      tsdbTbDataIterCreate(pTbData, NULL, 0, &pCommitIter->pIter);
 
       pCommitIter->pTable = (STable *)taosMemoryMalloc(sizeof(STable));
       pCommitIter->pTable->uid = pTbData->uid;
-      pCommitIter->pTable->tid = pTbData->uid;
+      pCommitIter->pTable->suid = pTbData->suid;
       pCommitIter->pTable->pSchema = pTSchema;
       pCommitIter->pTable->pCacheSchema = NULL;
     }
   }
-  tSkipListDestroyIter(pSlIter);
 
-  return 0;
+  return code;
+
+_err:
+  return code;
 }
 
 static void tsdbDestroyCommitIters(SCommitH *pCommith) {
   if (pCommith->iters == NULL) return;
 
   for (int i = 1; i < pCommith->niters; i++) {
-    tSkipListDestroyIter(pCommith->iters[i].pIter);
+    tsdbTbDataIterDestroy(pCommith->iters[i].pIter);
     if (pCommith->iters[i].pTable) {
       tdFreeSchema(pCommith->iters[i].pTable->pSchema);
       tdFreeSchema(pCommith->iters[i].pTable->pCacheSchema);
@@ -743,8 +734,8 @@ static int tsdbWriteBlockInfoImpl(SDFile *pHeadf, STable *pTable, SArray *pSupA,
   pBlkInfo = *ppBuf;
 
   pBlkInfo->delimiter = TSDB_FILE_DELIMITER;
-  pBlkInfo->tid = TABLE_TID(pTable);
-  pBlkInfo->uid = TABLE_UID(pTable);
+  pBlkInfo->suid = pTable->suid;
+  pBlkInfo->uid = pTable->uid;
 
   memcpy((void *)(pBlkInfo->blocks), taosArrayGet(pSupA, 0), nSupBlocks * sizeof(SBlock));
   if (nSubBlocks > 0) {
@@ -770,7 +761,8 @@ static int tsdbWriteBlockInfoImpl(SDFile *pHeadf, STable *pTable, SArray *pSupA,
   // Set pIdx
   pBlock = taosArrayGetLast(pSupA);
 
-  pIdx->uid = TABLE_UID(pTable);
+  pIdx->suid = pTable->suid;
+  pIdx->uid = pTable->uid;
   pIdx->hasLast = pBlock->last ? 1 : 0;
   pIdx->maxKey = pBlock->maxKey;
   pIdx->numOfBlocks = (uint32_t)nSupBlocks;
@@ -925,7 +917,7 @@ static int tsdbMoveBlkIdx(SCommitH *pCommith, SBlockIdx *pIdx) {
     return -1;
   }
 
-  STable table = {.tid = pIdx->uid, .uid = pIdx->uid, .pSchema = NULL};
+  STable table = {.suid = pIdx->suid, .uid = pIdx->uid, .pSchema = NULL};
   pCommith->pTable = &table;
 
   while (bidx < nBlocks) {
@@ -1186,7 +1178,7 @@ static int tsdbWriteBlockImpl(STsdb *pRepo, STable *pTable, SDFile *pDFile, SDFi
   }
 
   pBlockData->delimiter = TSDB_FILE_DELIMITER;
-  pBlockData->uid = TABLE_UID(pTable);
+  pBlockData->uid = pTable->uid;
   pBlockData->numOfCols = nColsNotAllNull;
 
   taosCalcChecksumAppend(0, (uint8_t *)pBlockData, tsize);
@@ -1226,7 +1218,7 @@ static int tsdbWriteBlockImpl(STsdb *pRepo, STable *pTable, SDFile *pDFile, SDFi
 
   tsdbDebug("vgId:%d, uid:%" PRId64 " a block of data is written to file %s, offset %" PRId64
             " numOfRows %d len %d numOfCols %" PRId16 " keyFirst %" PRId64 " keyLast %" PRId64,
-            REPO_ID(pRepo), TABLE_UID(pTable), TSDB_FILE_FULL_NAME(pDFile), offset, rowsToWrite, pBlock->len,
+            REPO_ID(pRepo), pTable->uid, TSDB_FILE_FULL_NAME(pDFile), offset, rowsToWrite, pBlock->len,
             pBlock->numOfCols, pBlock->minKey.ts, pBlock->maxKey.ts);
 
   return 0;
@@ -1313,7 +1305,7 @@ static int tsdbMergeMemData(SCommitH *pCommith, SCommitIter *pIter, int bidx) {
     keyLimit = pBlock[1].minKey.ts - 1;
   }
 
-  SSkipListIterator titer = *(pIter->pIter);
+  STbDataIter titer = *(pIter->pIter);
   if (tsdbLoadBlockDataCols(&(pCommith->readh), pBlock, NULL, &colId, 1, false) < 0) return -1;
 
   tsdbLoadDataFromCache(TSDB_COMMIT_REPO(pCommith), pIter->pTable, &titer, keyLimit, INT32_MAX, NULL,
@@ -1522,7 +1514,7 @@ static void tsdbLoadAndMergeFromCache(STsdb *pTsdb, SDataCols *pDataCols, int *i
         lastKey = key2;
       }
 
-      tSkipListIterNext(pCommitIter->pIter);
+      tsdbTbDataIterNext(pCommitIter->pIter);
     } else {
       if (lastKey != key1) {
         if (lastKey != TSKEY_INITIAL_VAL) {
@@ -1554,7 +1546,7 @@ static void tsdbLoadAndMergeFromCache(STsdb *pTsdb, SDataCols *pDataCols, int *i
         tdAppendSTSRowToDataCol(row, pSchema, pTarget, true);
       }
       ++(*iter);
-      tSkipListIterNext(pCommitIter->pIter);
+      tsdbTbDataIterNext(pCommitIter->pIter);
     }
 
     if (pTarget->numOfRows >= (maxRows - 1)) break;

source/dnode/vnode/src/tsdb/tsdbCommit2.c

@@ -16,19 +16,19 @@
 #include "tsdb.h"
 
 typedef struct {
-  SMemTable *pMemTable;
-  int32_t    minutes;
-  int8_t     precision;
-  TSKEY      nCommitKey;
-  int32_t    fid;
-  TSKEY      minKey;
-  TSKEY      maxKey;
-  SReadH     readh;
-  SDFileSet  wSet;
-  SArray    *aBlkIdx;
-  SArray    *aSupBlk;
-  SArray    *aSubBlk;
-  SArray    *aDelInfo;
+  SMemTable2 *pMemTable;
+  int32_t     minutes;
+  int8_t      precision;
+  TSKEY       nCommitKey;
+  int32_t     fid;
+  TSKEY       minKey;
+  TSKEY       maxKey;
+  SReadH      readh;
+  SDFileSet   wSet;
+  SArray     *aBlkIdx;
+  SArray     *aSupBlk;
+  SArray     *aSubBlk;
+  SArray     *aDelInfo;
 } SCommitH;
 
 static int32_t tsdbCommitStart(SCommitH *pCHandle, STsdb *pTsdb);
@@ -39,7 +39,7 @@ int32_t tsdbBegin2(STsdb *pTsdb) {
   int32_t code = 0;
 
   ASSERT(pTsdb->mem == NULL);
-  code = tsdbMemTableCreate2(pTsdb, (SMemTable **)&pTsdb->mem);
+  code = tsdbMemTableCreate2(pTsdb, (SMemTable2 **)&pTsdb->mem);
   if (code) {
     tsdbError("vgId:%d failed to begin TSDB since %s", TD_VID(pTsdb->pVnode), tstrerror(code));
     goto _exit;
@@ -80,8 +80,8 @@ _err:
 }
 
 static int32_t tsdbCommitStart(SCommitH *pCHandle, STsdb *pTsdb) {
-  int32_t    code = 0;
-  SMemTable *pMemTable = (SMemTable *)pTsdb->mem;
+  int32_t     code = 0;
+  SMemTable2 *pMemTable = (SMemTable2 *)pTsdb->mem;
 
   tsdbInfo("vgId:%d start to commit", TD_VID(pTsdb->pVnode));
 
@@ -131,9 +131,9 @@ _err:
 }
 
 static int32_t tsdbCommitEnd(SCommitH *pCHandle) {
-  int32_t    code = 0;
-  STsdb     *pTsdb = pCHandle->pMemTable->pTsdb;
-  SMemTable *pMemTable = (SMemTable *)pTsdb->imem;
+  int32_t     code = 0;
+  STsdb      *pTsdb = pCHandle->pMemTable->pTsdb;
+  SMemTable2 *pMemTable = (SMemTable2 *)pTsdb->imem;
 
   // end transaction
   code = tsdbEndFSTxn(pTsdb);

source/dnode/vnode/src/tsdb/tsdbFile.c

@@ -25,8 +25,7 @@ static const char *TSDB_FNAME_SUFFIX[] = {
     "meta",  // TSDB_FILE_META
 };
 
-static void tsdbGetFilename(int vid, int fid, uint32_t ver, TSDB_FILE_T ftype, const char *dname, char *fname);
-// static int   tsdbRollBackMFile(SMFile *pMFile);
+static void  tsdbGetFilename(int vid, int fid, uint32_t ver, TSDB_FILE_T ftype, const char *dname, char *fname);
 static int   tsdbEncodeDFInfo(void **buf, SDFInfo *pInfo);
 static void *tsdbDecodeDFInfo(void *buf, SDFInfo *pInfo);
 static int   tsdbRollBackDFile(SDFile *pDFile);

source/dnode/vnode/src/tsdb/tsdbMemTable2.c

@@ -15,11 +15,6 @@
 
 #include "tsdb.h"
 
-struct SMemSkipListNode {
-  int8_t            level;
-  SMemSkipListNode *forwards[0];
-};
-
 typedef struct {
   tb_uid_t  uid;
   STSchema *pTSchema;
@@ -35,21 +30,21 @@ typedef struct {
 #define SL_MOVE_BACKWARD 0x1
 #define SL_MOVE_FROM_POS 0x2
 
-static int32_t tsdbGetOrCreateMemData(SMemTable *pMemTable, tb_uid_t suid, tb_uid_t uid, SMemData **ppMemData);
+static int32_t tsdbGetOrCreateMemData(SMemTable2 *pMemTable, tb_uid_t suid, tb_uid_t uid, SMemData **ppMemData);
 static int     memDataPCmprFn(const void *p1, const void *p2);
 static int32_t tPutTSDBRow(uint8_t *p, TSDBROW *pRow);
 static int32_t tGetTSDBRow(uint8_t *p, TSDBROW *pRow);
 static int8_t  tsdbMemSkipListRandLevel(SMemSkipList *pSl);
-static int32_t tsdbInsertTableDataImpl(SMemTable *pMemTable, SMemData *pMemData, int64_t version,
+static int32_t tsdbInsertTableDataImpl(SMemTable2 *pMemTable, SMemData *pMemData, int64_t version,
                                        SVSubmitBlk *pSubmitBlk);
 static void    memDataMovePosTo(SMemData *pMemData, SMemSkipListNode **pos, TSDBKEY *pKey, int32_t flags);
 
 // SMemTable ==============================================
-int32_t tsdbMemTableCreate2(STsdb *pTsdb, SMemTable **ppMemTable) {
-  int32_t    code = 0;
-  SMemTable *pMemTable = NULL;
+int32_t tsdbMemTableCreate2(STsdb *pTsdb, SMemTable2 **ppMemTable) {
+  int32_t     code = 0;
+  SMemTable2 *pMemTable = NULL;
 
-  pMemTable = (SMemTable *)taosMemoryCalloc(1, sizeof(*pMemTable));
+  pMemTable = (SMemTable2 *)taosMemoryCalloc(1, sizeof(*pMemTable));
   if (pMemTable == NULL) {
     code = TSDB_CODE_OUT_OF_MEMORY;
     goto _err;
@@ -75,16 +70,16 @@ _err:
   return code;
 }
 
-void tsdbMemTableDestroy2(SMemTable *pMemTable) {
+void tsdbMemTableDestroy2(SMemTable2 *pMemTable) {
   taosArrayDestroyEx(pMemTable->aMemData, NULL /*TODO*/);
   taosMemoryFree(pMemTable);
 }
 
 int32_t tsdbInsertTableData2(STsdb *pTsdb, int64_t version, SVSubmitBlk *pSubmitBlk) {
-  int32_t    code = 0;
-  SMemTable *pMemTable = (SMemTable *)pTsdb->mem;  // TODO
-  SMemData  *pMemData;
-  TSDBROW    row = {.version = version};
+  int32_t     code = 0;
+  SMemTable2 *pMemTable = (SMemTable2 *)pTsdb->mem;  // TODO
+  SMemData   *pMemData;
+  TSDBROW     row = {.version = version};
 
   ASSERT(pMemTable);
   ASSERT(pSubmitBlk->nData > 0);
@@ -112,10 +107,10 @@ _err:
 }
 
 int32_t tsdbDeleteTableData2(STsdb *pTsdb, int64_t version, tb_uid_t suid, tb_uid_t uid, TSKEY sKey, TSKEY eKey) {
-  int32_t    code = 0;
-  SMemTable *pMemTable = (SMemTable *)pTsdb->mem;  // TODO
-  SMemData  *pMemData;
-  SVBufPool *pPool = pTsdb->pVnode->inUse;
+  int32_t     code = 0;
+  SMemTable2 *pMemTable = (SMemTable2 *)pTsdb->mem;  // TODO
+  SMemData   *pMemData;
+  SVBufPool  *pPool = pTsdb->pVnode->inUse;
 
   ASSERT(pMemTable);
 
@@ -250,7 +245,7 @@ void tsdbMemDataIterGet(SMemDataIter *pIter, TSDBROW **ppRow) {
   }
 }
 
-static int32_t tsdbGetOrCreateMemData(SMemTable *pMemTable, tb_uid_t suid, tb_uid_t uid, SMemData **ppMemData) {
+static int32_t tsdbGetOrCreateMemData(SMemTable2 *pMemTable, tb_uid_t suid, tb_uid_t uid, SMemData **ppMemData) {
   int32_t    code = 0;
   int32_t    idx = 0;
   SMemData  *pMemDataT = &(SMemData){.suid = suid, .uid = uid};
@@ -421,7 +416,7 @@ static void memDataMovePosTo(SMemData *pMemData, SMemSkipListNode **pos, TSDBKEY
   }
 }
 
-static int32_t memDataDoPut(SMemTable *pMemTable, SMemData *pMemData, SMemSkipListNode **pos, TSDBROW *pRow,
+static int32_t memDataDoPut(SMemTable2 *pMemTable, SMemData *pMemData, SMemSkipListNode **pos, TSDBROW *pRow,
                             int8_t forward) {
   int32_t           code = 0;
   int8_t            level;
@@ -475,7 +470,7 @@ _exit:
   return code;
 }
 
-static int32_t tsdbInsertTableDataImpl(SMemTable *pMemTable, SMemData *pMemData, int64_t version,
+static int32_t tsdbInsertTableDataImpl(SMemTable2 *pMemTable, SMemData *pMemData, int64_t version,
                                        SVSubmitBlk *pSubmitBlk) {
   int32_t  code = 0;
   int32_t  n = 0;

source/dnode/vnode/src/tsdb/tsdbRead.c

@@ -67,15 +67,16 @@ enum {
 };
 
 typedef struct STableCheckInfo {
-  uint64_t           tableId;
-  TSKEY              lastKey;
-  SBlockInfo*        pCompInfo;
-  int32_t            compSize;
-  int32_t            numOfBlocks : 29;  // number of qualified data blocks not the original blocks
-  uint8_t            chosen : 2;        // indicate which iterator should move forward
-  bool               initBuf : 1;       // whether to initialize the in-memory skip list iterator or not
-  SSkipListIterator* iter;              // mem buffer skip list iterator
-  SSkipListIterator* iiter;             // imem buffer skip list iterator
+  uint64_t     suid;
+  uint64_t     tableId;
+  TSKEY        lastKey;
+  SBlockInfo*  pCompInfo;
+  int32_t      compSize;
+  int32_t      numOfBlocks : 29;  // number of qualified data blocks not the original blocks
+  uint8_t      chosen : 2;        // indicate which iterator should move forward
+  bool         initBuf : 1;       // whether to initialize the in-memory skip list iterator or not
+  STbDataIter* iter;              // mem buffer skip list iterator
+  STbDataIter* iiter;             // imem buffer skip list iterator
 } STableCheckInfo;
 
 typedef struct STableBlockInfo {
@@ -107,6 +108,7 @@ typedef struct SBlockLoadSuppInfo {
 
 typedef struct STsdbReadHandle {
   STsdb*        pTsdb;
+  uint64_t      suid;
   SQueryFilePos cur;  // current position
   int16_t       order;
   STimeWindow   window;  // the primary query time window that applies to all queries
@@ -200,8 +202,8 @@ static SArray* getDefaultLoadColumns(STsdbReadHandle* pTsdbReadHandle, bool load
 int64_t tsdbGetNumOfRowsInMemTable(tsdbReaderT* pHandle) {
   STsdbReadHandle* pTsdbReadHandle = (STsdbReadHandle*)pHandle;
 
-  int64_t        rows = 0;
-  STsdbMemTable* pMemTable = NULL;  // pTsdbReadHandle->pMemTable;
+  int64_t    rows = 0;
+  SMemTable* pMemTable = NULL;  // pTsdbReadHandle->pMemTable;
   if (pMemTable == NULL) {
     return rows;
   }
@@ -237,6 +239,7 @@ static SArray* createCheckInfoFromTableGroup(STsdbReadHandle* pTsdbReadHandle, S
     STableKeyInfo* pKeyInfo = (STableKeyInfo*)taosArrayGet(pTableList->pTableList, j);
 
     STableCheckInfo info = {.lastKey = pKeyInfo->lastKey, .tableId = pKeyInfo->uid};
+    info.suid = pTsdbReadHandle->suid;
     if (ASCENDING_TRAVERSE(pTsdbReadHandle->order)) {
       if (info.lastKey == INT64_MIN || info.lastKey < pTsdbReadHandle->window.skey) {
         info.lastKey = pTsdbReadHandle->window.skey;
@@ -265,8 +268,8 @@ static void resetCheckInfo(STsdbReadHandle* pTsdbReadHandle) {
   for (int32_t i = 0; i < numOfTables; ++i) {
     STableCheckInfo* pCheckInfo = (STableCheckInfo*)taosArrayGet(pTsdbReadHandle->pTableCheckInfo, i);
     pCheckInfo->lastKey = pTsdbReadHandle->window.skey;
-    pCheckInfo->iter = tSkipListDestroyIter(pCheckInfo->iter);
-    pCheckInfo->iiter = tSkipListDestroyIter(pCheckInfo->iiter);
+    pCheckInfo->iter = tsdbTbDataIterDestroy(pCheckInfo->iter);
+    pCheckInfo->iiter = tsdbTbDataIterDestroy(pCheckInfo->iiter);
     pCheckInfo->initBuf = false;
 
     if (ASCENDING_TRAVERSE(pTsdbReadHandle->order)) {
@@ -387,6 +390,7 @@ static STsdbReadHandle* tsdbQueryTablesImpl(SVnode* pVnode, SQueryTableDataCond*
   pReadHandle->locateStart = false;
   pReadHandle->loadType = pCond->type;
 
+  pReadHandle->suid = pCond->suid;
   pReadHandle->outputCapacity = 4096;  //((STsdb*)tsdb)->config.maxRowsPerFileBlock;
   pReadHandle->loadExternalRow = pCond->loadExternalRows;
   pReadHandle->currentLoadExternalRows = pCond->loadExternalRows;
@@ -658,7 +662,7 @@ tsdbReaderT tsdbQueryLastRow(SVnode* pVnode, SQueryTableDataCond* pCond, STableL
 }
 
 #if 0
-tsdbReaderT tsdbQueryCacheLastT(STsdb *tsdb, SQueryTableDataCond *pCond, STableGroupInfo *groupList, uint64_t qId, STsdbMemTable* pMemRef) {
+tsdbReaderT tsdbQueryCacheLastT(STsdb *tsdb, SQueryTableDataCond *pCond, STableGroupInfo *groupList, uint64_t qId, SMemTable* pMemRef) {
   STsdbReadHandle *pTsdbReadHandle = (STsdbReadHandle*) tsdbQueryTablesT(tsdb, pCond, groupList, qId, pMemRef);
   if (pTsdbReadHandle == NULL) {
     return NULL;
@@ -752,23 +756,22 @@ static bool initTableMemIterator(STsdbReadHandle* pHandle, STableCheckInfo* pChe
   pCheckInfo->initBuf = true;
   int32_t order = pHandle->order;
 
-  STbData** pMem = NULL;
-  STbData** pIMem = NULL;
+  STbData* pMem = NULL;
+  STbData* pIMem = NULL;
+  int8_t   backward = (pHandle->order == TSDB_ORDER_DESC) ? 1 : 0;
 
   TSKEY tLastKey = keyToTkey(pCheckInfo->lastKey);
   if (pHandle->pTsdb->mem != NULL) {
-    pMem = taosHashGet(pHandle->pTsdb->mem->pHashIdx, &pCheckInfo->tableId, sizeof(pCheckInfo->tableId));
+    tsdbGetTbDataFromMemTable(pHandle->pTsdb->mem, pCheckInfo->suid, pCheckInfo->tableId, &pMem);
     if (pMem != NULL) {
-      pCheckInfo->iter =
-          tSkipListCreateIterFromVal((*pMem)->pData, (const char*)&tLastKey, TSDB_DATA_TYPE_TIMESTAMP, order);
+      tsdbTbDataIterCreate(pMem, &(TSDBKEY){.version = 0, .ts = tLastKey}, backward, &pCheckInfo->iter);
     }
   }
 
   if (pHandle->pTsdb->imem != NULL) {
-    pIMem = taosHashGet(pHandle->pTsdb->imem->pHashIdx, &pCheckInfo->tableId, sizeof(pCheckInfo->tableId));
+    tsdbGetTbDataFromMemTable(pHandle->pTsdb->mem, pCheckInfo->suid, pCheckInfo->tableId, &pIMem);
     if (pIMem != NULL) {
-      pCheckInfo->iiter =
-          tSkipListCreateIterFromVal((*pIMem)->pData, (const char*)&tLastKey, TSDB_DATA_TYPE_TIMESTAMP, order);
+      tsdbTbDataIterCreate(pIMem, &(TSDBKEY){.version = 0, .ts = tLastKey}, backward, &pCheckInfo->iiter);
     }
   }
 
@@ -777,22 +780,23 @@ static bool initTableMemIterator(STsdbReadHandle* pHandle, STableCheckInfo* pChe
     return false;
   }
 
-  bool memEmpty = (pCheckInfo->iter == NULL) || (pCheckInfo->iter != NULL && !tSkipListIterNext(pCheckInfo->iter));
-  bool imemEmpty = (pCheckInfo->iiter == NULL) || (pCheckInfo->iiter != NULL && !tSkipListIterNext(pCheckInfo->iiter));
+  bool memEmpty =
+      (pCheckInfo->iter == NULL) || (pCheckInfo->iter != NULL && !tsdbTbDataIterGet(pCheckInfo->iter, NULL));
+  bool imemEmpty =
+      (pCheckInfo->iiter == NULL) || (pCheckInfo->iiter != NULL && !tsdbTbDataIterGet(pCheckInfo->iiter, NULL));
   if (memEmpty && imemEmpty) {  // buffer is empty
     return false;
   }
 
   if (!memEmpty) {
-    SSkipListNode* node = tSkipListIterGet(pCheckInfo->iter);
-    assert(node != NULL);
+    TSDBROW row;
 
-    STSRow* row = (STSRow*)SL_GET_NODE_DATA(node);
-    TSKEY   key = TD_ROW_KEY(row);  // first timestamp in buffer
+    tsdbTbDataIterGet(pCheckInfo->iter, &row);
+    TSKEY key = row.pTSRow->ts;  // first timestamp in buffer
     tsdbDebug("%p uid:%" PRId64 ", check data in mem from skey:%" PRId64 ", order:%d, ts range in buf:%" PRId64
               "-%" PRId64 ", lastKey:%" PRId64 ", numOfRows:%" PRId64 ", %s",
-              pHandle, pCheckInfo->tableId, key, order, (*pMem)->keyMin, (*pMem)->keyMax, pCheckInfo->lastKey,
-              (*pMem)->nrows, pHandle->idStr);
+              pHandle, pCheckInfo->tableId, key, order, pMem->minKey.ts, pMem->maxKey.ts, pCheckInfo->lastKey,
+              pMem->sl.size, pHandle->idStr);
 
     if (ASCENDING_TRAVERSE(order)) {
       assert(pCheckInfo->lastKey <= key);
@@ -805,15 +809,14 @@ static bool initTableMemIterator(STsdbReadHandle* pHandle, STableCheckInfo* pChe
   }
 
   if (!imemEmpty) {
-    SSkipListNode* node = tSkipListIterGet(pCheckInfo->iiter);
-    assert(node != NULL);
+    TSDBROW row;
 
-    STSRow* row = (STSRow*)SL_GET_NODE_DATA(node);
-    TSKEY   key = TD_ROW_KEY(row);  // first timestamp in buffer
+    tsdbTbDataIterGet(pCheckInfo->iter, &row);
+    TSKEY key = row.pTSRow->ts;  // first timestamp in buffer
     tsdbDebug("%p uid:%" PRId64 ", check data in imem from skey:%" PRId64 ", order:%d, ts range in buf:%" PRId64
               "-%" PRId64 ", lastKey:%" PRId64 ", numOfRows:%" PRId64 ", %s",
-              pHandle, pCheckInfo->tableId, key, order, (*pIMem)->keyMin, (*pIMem)->keyMax, pCheckInfo->lastKey,
-              (*pIMem)->nrows, pHandle->idStr);
+              pHandle, pCheckInfo->tableId, key, order, pIMem->minKey.ts, pIMem->maxKey.ts, pCheckInfo->lastKey,
+              pIMem->sl.size, pHandle->idStr);
 
     if (ASCENDING_TRAVERSE(order)) {
       assert(pCheckInfo->lastKey <= key);
@@ -828,31 +831,23 @@ static bool initTableMemIterator(STsdbReadHandle* pHandle, STableCheckInfo* pChe
 }
 
 static void destroyTableMemIterator(STableCheckInfo* pCheckInfo) {
-  tSkipListDestroyIter(pCheckInfo->iter);
-  tSkipListDestroyIter(pCheckInfo->iiter);
+  tsdbTbDataIterDestroy(pCheckInfo->iter);
+  tsdbTbDataIterDestroy(pCheckInfo->iiter);
 }
 
 static TSKEY extractFirstTraverseKey(STableCheckInfo* pCheckInfo, int32_t order, int32_t update, TDRowVerT maxVer) {
+  TSDBROW row = {0};
   STSRow *rmem = NULL, *rimem = NULL;
+
   if (pCheckInfo->iter) {
-    SSkipListNode* node = tSkipListIterGet(pCheckInfo->iter);
-    if (node != NULL) {
-      rmem = (STSRow*)SL_GET_NODE_DATA(node);
-      // TODO: filter max version
-      // if (TD_ROW_VER(rmem) > maxVer) {
-      //   rmem = NULL;
-      // }
+    if (tsdbTbDataIterGet(pCheckInfo->iter, &row)) {
+      rmem = row.pTSRow;
     }
   }
 
   if (pCheckInfo->iiter) {
-    SSkipListNode* node = tSkipListIterGet(pCheckInfo->iiter);
-    if (node != NULL) {
-      rimem = (STSRow*)SL_GET_NODE_DATA(node);
-      // TODO: filter max version
-      // if (TD_ROW_VER(rimem) > maxVer) {
-      //   rimem = NULL;
-      // }
+    if (tsdbTbDataIterGet(pCheckInfo->iiter, &row)) {
+      rimem = row.pTSRow;
     }
   }
 
@@ -889,7 +884,7 @@ static TSKEY extractFirstTraverseKey(STableCheckInfo* pCheckInfo, int32_t order,
       pCheckInfo->chosen = CHECKINFO_CHOSEN_BOTH;
     } else {
       pCheckInfo->chosen = CHECKINFO_CHOSEN_IMEM;
-      tSkipListIterNext(pCheckInfo->iter);
+      tsdbTbDataIterNext(pCheckInfo->iter);
     }
     return r1;
   } else if (r1 < r2 && ASCENDING_TRAVERSE(order)) {
@@ -903,28 +898,17 @@ static TSKEY extractFirstTraverseKey(STableCheckInfo* pCheckInfo, int32_t order,
 
 static STSRow* getSRowInTableMem(STableCheckInfo* pCheckInfo, int32_t order, int32_t update, STSRow** extraRow,
                                  TDRowVerT maxVer) {
+  TSDBROW row;
   STSRow *rmem = NULL, *rimem = NULL;
   if (pCheckInfo->iter) {
-    SSkipListNode* node = tSkipListIterGet(pCheckInfo->iter);
-    if (node != NULL) {
-      rmem = (STSRow*)SL_GET_NODE_DATA(node);
-#if 0  // TODO: skiplist refactor
-      if (TD_ROW_VER(rmem) > maxVer) {
-        rmem = NULL;
-      }
-#endif
+    if (tsdbTbDataIterGet(pCheckInfo->iter, &row)) {
+      rmem = row.pTSRow;
     }
   }
 
   if (pCheckInfo->iiter) {
-    SSkipListNode* node = tSkipListIterGet(pCheckInfo->iiter);
-    if (node != NULL) {
-      rimem = (STSRow*)SL_GET_NODE_DATA(node);
-#if 0  // TODO: skiplist refactor
-      if (TD_ROW_VER(rimem) > maxVer) {
-        rimem = NULL;
-      }
-#endif
+    if (tsdbTbDataIterGet(pCheckInfo->iiter, &row)) {
+      rimem = row.pTSRow;
     }
   }
 
@@ -966,7 +950,7 @@ static STSRow* getSRowInTableMem(STableCheckInfo* pCheckInfo, int32_t order, int
       *extraRow = rimem;
       return rmem;
     } else {
-      tSkipListIterNext(pCheckInfo->iter);
+      tsdbTbDataIterNext(pCheckInfo->iter);
       pCheckInfo->chosen = CHECKINFO_CHOSEN_IMEM;
       return rimem;
     }
@@ -995,7 +979,7 @@ static bool moveToNextRowInMem(STableCheckInfo* pCheckInfo) {
   bool hasNext = false;
   if (pCheckInfo->chosen == CHECKINFO_CHOSEN_MEM) {
     if (pCheckInfo->iter != NULL) {
-      hasNext = tSkipListIterNext(pCheckInfo->iter);
+      hasNext = tsdbTbDataIterNext(pCheckInfo->iter);
     }
 
     if (hasNext) {
@@ -1003,11 +987,11 @@ static bool moveToNextRowInMem(STableCheckInfo* pCheckInfo) {
     }
 
     if (pCheckInfo->iiter != NULL) {
-      return tSkipListIterGet(pCheckInfo->iiter) != NULL;
+      return tsdbTbDataIterGet(pCheckInfo->iiter, NULL);
     }
   } else if (pCheckInfo->chosen == CHECKINFO_CHOSEN_IMEM) {
     if (pCheckInfo->iiter != NULL) {
-      hasNext = tSkipListIterNext(pCheckInfo->iiter);
+      hasNext = tsdbTbDataIterNext(pCheckInfo->iiter);
     }
 
     if (hasNext) {
@@ -1015,14 +999,14 @@ static bool moveToNextRowInMem(STableCheckInfo* pCheckInfo) {
     }
 
     if (pCheckInfo->iter != NULL) {
-      return tSkipListIterGet(pCheckInfo->iter) != NULL;
+      return tsdbTbDataIterGet(pCheckInfo->iter, NULL);
     }
   } else {
     if (pCheckInfo->iter != NULL) {
-      hasNext = tSkipListIterNext(pCheckInfo->iter);
+      hasNext = tsdbTbDataIterNext(pCheckInfo->iter);
     }
     if (pCheckInfo->iiter != NULL) {
-      hasNext = tSkipListIterNext(pCheckInfo->iiter) || hasNext;
+      hasNext = tsdbTbDataIterNext(pCheckInfo->iiter) || hasNext;
     }
   }
 
@@ -1126,7 +1110,7 @@ static int32_t loadBlockInfo(STsdbReadHandle* pTsdbReadHandle, int32_t index, in
   STableCheckInfo* pCheckInfo = taosArrayGet(pTsdbReadHandle->pTableCheckInfo, index);
   pCheckInfo->numOfBlocks = 0;
 
-  STable table = {.uid = pCheckInfo->tableId, .tid = pCheckInfo->tableId};
+  STable table = {.uid = pCheckInfo->tableId, .suid = pCheckInfo->suid};
   table.pSchema = pTsdbReadHandle->pSchema;
 
   if (tsdbSetReadTable(&pTsdbReadHandle->rhelper, &table) != TSDB_CODE_SUCCESS) {
@@ -2924,7 +2908,7 @@ static bool loadBlockOfActiveTable(STsdbReadHandle* pTsdbReadHandle) {
   // current result is empty
   if (pTsdbReadHandle->currentLoadExternalRows && pTsdbReadHandle->window.skey == pTsdbReadHandle->window.ekey &&
       pTsdbReadHandle->cur.rows == 0) {
-    //    STsdbMemTable* pMemRef = pTsdbReadHandle->pMemTable;
+    //    SMemTable* pMemRef = pTsdbReadHandle->pMemTable;
 
     //    doGetExternalRow(pTsdbReadHandle, TSDB_PREV_ROW, pMemRef);
     //    doGetExternalRow(pTsdbReadHandle, TSDB_NEXT_ROW, pMemRef);
@@ -3222,7 +3206,7 @@ bool tsdbNextDataBlock(tsdbReaderT pHandle) {
   }
 }
 
-// static int32_t doGetExternalRow(STsdbReadHandle* pTsdbReadHandle, int16_t type, STsdbMemTable* pMemRef) {
+// static int32_t doGetExternalRow(STsdbReadHandle* pTsdbReadHandle, int16_t type, SMemTable* pMemRef) {
 //  STsdbReadHandle* pSecQueryHandle = NULL;
 //
 //  if (type == TSDB_PREV_ROW && pTsdbReadHandle->prev) {

source/dnode/vnode/src/tsdb/tsdbReadImpl.c

@@ -156,6 +156,24 @@ int tsdbLoadBlockIdx(SReadH *pReadh) {
   return 0;
 }
 
+static int32_t tsdbBlockIdxCmprFn(const void *p1, const void *p2) {
+  SBlockIdx *pBlockIdx1 = (SBlockIdx *)p1;
+  SBlockIdx *pBlockIdx2 = (SBlockIdx *)p2;
+
+  if (pBlockIdx1->suid < pBlockIdx2->suid) {
+    return -1;
+  } else if (pBlockIdx1->suid > pBlockIdx2->suid) {
+    return 1;
+  }
+
+  if (pBlockIdx1->uid < pBlockIdx2->uid) {
+    return -1;
+  } else if (pBlockIdx1->uid > pBlockIdx2->uid) {
+    return 1;
+  }
+
+  return 0;
+}
 int tsdbSetReadTable(SReadH *pReadh, STable *pTable) {
   STSchema *pSchema = tsdbGetTableSchemaImpl(TSDB_READ_REPO(pReadh), pTable, false, false, -1);
 
@@ -171,33 +189,40 @@ int tsdbSetReadTable(SReadH *pReadh, STable *pTable) {
     return -1;
   }
 
-  size_t size = taosArrayGetSize(pReadh->aBlkIdx);
-  if (size > 0) {
-    while (true) {
-      if (pReadh->cidx >= size) {
-        pReadh->pBlkIdx = NULL;
-        break;
-      }
-
-      SBlockIdx *pBlkIdx = taosArrayGet(pReadh->aBlkIdx, pReadh->cidx);
-      if (pBlkIdx->uid == TABLE_TID(pTable)) {
-        if (pBlkIdx->uid == TABLE_UID(pTable)) {
-          pReadh->pBlkIdx = pBlkIdx;
-        } else {
-          pReadh->pBlkIdx = NULL;
-        }
-        pReadh->cidx++;
-        break;
-      } else if (pBlkIdx->uid > TABLE_TID(pTable)) {
-        pReadh->pBlkIdx = NULL;
-        break;
-      } else {
-        pReadh->cidx++;
-      }
-    }
-  } else {
+  uint8_t *p = taosArraySearch(pReadh->aBlkIdx, &(SBlockIdx){.suid = pTable->suid, .uid = pTable->uid},
+                               tsdbBlockIdxCmprFn, TD_EQ);
+  if (p == NULL) {
     pReadh->pBlkIdx = NULL;
-  }
+  } else {
+    pReadh->pBlkIdx = (SBlockIdx *)p;
+  }
+  // size_t size = taosArrayGetSize(pReadh->aBlkIdx);
+  // if (size > 0) {
+  //   while (true) {
+  //     if (pReadh->cidx >= size) {
+  //       pReadh->pBlkIdx = NULL;
+  //       break;
+  //     }
+
+  //     SBlockIdx *pBlkIdx = taosArrayGet(pReadh->aBlkIdx, pReadh->cidx);
+  //     if (pBlkIdx->uid == TABLE_TID(pTable)) {
+  //       if (pBlkIdx->uid == TABLE_UID(pTable)) {
+  //         pReadh->pBlkIdx = pBlkIdx;
+  //       } else {
+  //         pReadh->pBlkIdx = NULL;
+  //       }
+  //       pReadh->cidx++;
+  //       break;
+  //     } else if (pBlkIdx->uid > TABLE_TID(pTable)) {
+  //       pReadh->pBlkIdx = NULL;
+  //       break;
+  //     } else {
+  //       pReadh->cidx++;
+  //     }
+  //   }
+  // } else {
+  //   pReadh->pBlkIdx = NULL;
+  // }
 
   return 0;
 }
@@ -553,12 +578,12 @@ static int tsdbLoadBlockOffset(SReadH *pReadh, SBlock *pBlock) {
 int tsdbEncodeSBlockIdx(void **buf, SBlockIdx *pIdx) {
   int tlen = 0;
 
-  // tlen += taosEncodeVariantI32(buf, pIdx->tid);
+  tlen += taosEncodeFixedU64(buf, pIdx->suid);
+  tlen += taosEncodeFixedU64(buf, pIdx->uid);
   tlen += taosEncodeVariantU32(buf, pIdx->len);
   tlen += taosEncodeVariantU32(buf, pIdx->offset);
   tlen += taosEncodeFixedU8(buf, pIdx->hasLast);
   tlen += taosEncodeVariantU32(buf, pIdx->numOfBlocks);
-  tlen += taosEncodeFixedU64(buf, pIdx->uid);
   tlen += taosEncodeFixedU64(buf, pIdx->maxKey.ts);
 
   return tlen;
@@ -570,6 +595,10 @@ void *tsdbDecodeSBlockIdx(void *buf, SBlockIdx *pIdx) {
   uint64_t value = 0;
 
   // if ((buf = taosDecodeVariantI32(buf, &(pIdx->tid))) == NULL) return NULL;
+  if ((buf = taosDecodeFixedU64(buf, &value)) == NULL) return NULL;
+  pIdx->suid = (int64_t)value;
+  if ((buf = taosDecodeFixedU64(buf, &value)) == NULL) return NULL;
+  pIdx->uid = (int64_t)value;
   if ((buf = taosDecodeVariantU32(buf, &(pIdx->len))) == NULL) return NULL;
   if ((buf = taosDecodeVariantU32(buf, &(pIdx->offset))) == NULL) return NULL;
   if ((buf = taosDecodeFixedU8(buf, &(hasLast))) == NULL) return NULL;
@@ -577,8 +606,6 @@ void *tsdbDecodeSBlockIdx(void *buf, SBlockIdx *pIdx) {
   if ((buf = taosDecodeVariantU32(buf, &(numOfBlocks))) == NULL) return NULL;
   pIdx->numOfBlocks = numOfBlocks;
   if ((buf = taosDecodeFixedU64(buf, &value)) == NULL) return NULL;
-  pIdx->uid = (int64_t)value;
-  if ((buf = taosDecodeFixedU64(buf, &value)) == NULL) return NULL;
   pIdx->maxKey.ts = (TSKEY)value;
 
   return buf;

source/dnode/vnode/src/tsdb/tsdbWrite.c

@@ -39,7 +39,7 @@ int tsdbInsertData(STsdb *pTsdb, int64_t version, SSubmitReq *pMsg, SSubmitRsp *
     SSubmitBlkRsp r = {0};
     tGetSubmitMsgNext(&msgIter, &pBlock);
     if (pBlock == NULL) break;
-    if (tsdbInsertTableData(pTsdb, &msgIter, pBlock, &r) < 0) {
+    if (tsdbInsertTableData(pTsdb, version, &msgIter, pBlock, &r) < 0) {
       return -1;
     }
 

source/dnode/vnode/src/vnd/vnodeSvr.c

@@ -799,7 +799,7 @@ static int32_t vnodeProcessSubmitReq(SVnode *pVnode, int64_t version, void *pReq
       sprintf(submitBlkRsp.tblFName, "%s.", pVnode->config.dbname);
     }
 
-    if (tsdbInsertTableData(pVnode->pTsdb, &msgIter, pBlock, &submitBlkRsp) < 0) {
+    if (tsdbInsertTableData(pVnode->pTsdb, version, &msgIter, pBlock, &submitBlkRsp) < 0) {
       submitBlkRsp.code = terrno;
     }
 

source/libs/executor/src/executorimpl.c

@@ -4760,6 +4760,7 @@ int32_t initQueryTableDataCond(SQueryTableDataCond* pCond, const STableScanPhysi
   pCond->numOfTWindows = 1;
   pCond->twindows = taosMemoryCalloc(pCond->numOfTWindows, sizeof(STimeWindow));
   pCond->twindows[0] = pTableScanNode->scanRange;
+  pCond->suid = pTableScanNode->scan.suid;
 
 #if 1
   // todo work around a problem, remove it later

source/libs/nodes/src/nodesCloneFuncs.c

@@ -449,6 +449,34 @@ static SNode* logicSubplanCopy(const SLogicSubplan* pSrc, SLogicSubplan* pDst) {
   return (SNode*)pDst;
 }
 
+static SNode* physiNodeCopy(const SPhysiNode* pSrc, SPhysiNode* pDst) {
+  CLONE_NODE_FIELD(pOutputDataBlockDesc);
+  CLONE_NODE_FIELD(pConditions);
+  CLONE_NODE_LIST_FIELD(pChildren);
+  return (SNode*)pDst;
+}
+
+static SNode* physiWindowCopy(const SWinodwPhysiNode* pSrc, SWinodwPhysiNode* pDst) {
+  COPY_BASE_OBJECT_FIELD(node, physiNodeCopy);
+  CLONE_NODE_LIST_FIELD(pExprs);
+  CLONE_NODE_LIST_FIELD(pFuncs);
+  CLONE_NODE_FIELD(pTspk);
+  COPY_SCALAR_FIELD(triggerType);
+  COPY_SCALAR_FIELD(watermark);
+  COPY_SCALAR_FIELD(filesFactor);
+  return (SNode*)pDst;
+}
+
+static SNode* physiIntervalCopy(const SIntervalPhysiNode* pSrc, SIntervalPhysiNode* pDst) {
+  COPY_BASE_OBJECT_FIELD(window, physiWindowCopy);
+  COPY_SCALAR_FIELD(interval);
+  COPY_SCALAR_FIELD(offset);
+  COPY_SCALAR_FIELD(sliding);
+  COPY_SCALAR_FIELD(intervalUnit);
+  COPY_SCALAR_FIELD(slidingUnit);
+  return (SNode*)pDst;
+}
+
 static SNode* dataBlockDescCopy(const SDataBlockDescNode* pSrc, SDataBlockDescNode* pDst) {
   COPY_SCALAR_FIELD(dataBlockId);
   CLONE_NODE_LIST_FIELD(pSlots);
@@ -575,6 +603,12 @@ SNodeptr nodesCloneNode(const SNodeptr pNode) {
       return logicIndefRowsFuncCopy((const SIndefRowsFuncLogicNode*)pNode, (SIndefRowsFuncLogicNode*)pDst);
     case QUERY_NODE_LOGIC_SUBPLAN:
       return logicSubplanCopy((const SLogicSubplan*)pNode, (SLogicSubplan*)pDst);
+    case QUERY_NODE_PHYSICAL_PLAN_HASH_INTERVAL:
+    case QUERY_NODE_PHYSICAL_PLAN_MERGE_INTERVAL:
+    case QUERY_NODE_PHYSICAL_PLAN_STREAM_INTERVAL:
+    case QUERY_NODE_PHYSICAL_PLAN_STREAM_FINAL_INTERVAL:
+    case QUERY_NODE_PHYSICAL_PLAN_STREAM_SEMI_INTERVAL:
+      return physiIntervalCopy((const SIntervalPhysiNode*)pNode, (SIntervalPhysiNode*)pDst);
     default:
       break;
   }

source/libs/nodes/src/nodesUtilFuncs.c

@@ -164,9 +164,14 @@ SNodeptr nodesMakeNode(ENodeType type) {
       return makeNode(type, sizeof(SCreateStreamStmt));
     case QUERY_NODE_DROP_STREAM_STMT:
       return makeNode(type, sizeof(SDropStreamStmt));
+    case QUERY_NODE_BALANCE_VGROUP_STMT:
+      return makeNode(type, sizeof(SBalanceVgroupStmt));
     case QUERY_NODE_MERGE_VGROUP_STMT:
+      return makeNode(type, sizeof(SMergeVgroupStmt));
     case QUERY_NODE_REDISTRIBUTE_VGROUP_STMT:
+      return makeNode(type, sizeof(SRedistributeVgroupStmt));
     case QUERY_NODE_SPLIT_VGROUP_STMT:
+      return makeNode(type, sizeof(SSplitVgroupStmt));
     case QUERY_NODE_SYNCDB_STMT:
       break;
     case QUERY_NODE_GRANT_STMT:

source/libs/parser/inc/parAst.h

@@ -187,6 +187,7 @@ SNode* createCreateStreamStmt(SAstCreateContext* pCxt, bool ignoreExists, const
                               SNode* pOptions, SNode* pQuery);
 SNode* createDropStreamStmt(SAstCreateContext* pCxt, bool ignoreNotExists, const SToken* pStreamName);
 SNode* createKillStmt(SAstCreateContext* pCxt, ENodeType type, const SToken* pId);
+SNode* createBalanceVgroupStmt(SAstCreateContext* pCxt);
 SNode* createMergeVgroupStmt(SAstCreateContext* pCxt, const SToken* pVgId1, const SToken* pVgId2);
 SNode* createRedistributeVgroupStmt(SAstCreateContext* pCxt, const SToken* pVgId, SNodeList* pDnodes);
 SNode* createSplitVgroupStmt(SAstCreateContext* pCxt, const SToken* pVgId);

source/libs/parser/inc/parInsertData.h

@@ -116,8 +116,7 @@ static FORCE_INLINE void getSTSRowAppendInfo(uint8_t rowType, SParsedDataColInfo
 }
 
 static FORCE_INLINE int32_t setBlockInfo(SSubmitBlk *pBlocks, STableDataBlocks *dataBuf, int32_t numOfRows) {
-  pBlocks->suid =
-      (TSDB_NORMAL_TABLE == dataBuf->pTableMeta->tableType ? dataBuf->pTableMeta->uid : dataBuf->pTableMeta->suid);
+  pBlocks->suid = (TSDB_NORMAL_TABLE == dataBuf->pTableMeta->tableType ? 0 : dataBuf->pTableMeta->suid);
   pBlocks->uid = dataBuf->pTableMeta->uid;
   pBlocks->sversion = dataBuf->pTableMeta->sversion;
   pBlocks->schemaLen = dataBuf->createTbReqLen;

source/libs/parser/inc/sql.y

@@ -469,9 +469,10 @@ cmd ::= KILL QUERY NK_INTEGER(A).
 cmd ::= KILL TRANSACTION NK_INTEGER(A).                                           { pCxt->pRootNode = createKillStmt(pCxt, QUERY_NODE_KILL_TRANSACTION_STMT, &A); }
 
 /************************************************ merge/redistribute/ vgroup ******************************************/
+cmd ::= BALANCE VGROUP.                                                           { pCxt->pRootNode = createBalanceVgroupStmt(pCxt); }
 cmd ::= MERGE VGROUP NK_INTEGER(A) NK_INTEGER(B).                                 { pCxt->pRootNode = createMergeVgroupStmt(pCxt, &A, &B); }
 cmd ::= REDISTRIBUTE VGROUP NK_INTEGER(A) dnode_list(B).                          { pCxt->pRootNode = createRedistributeVgroupStmt(pCxt, &A, B); }
-cmd ::= SPLIT VGROUP NK_INTEGER(A).                                               { pCxt->pRootNode = createSplitVgroupStmt(pCxt, &A); }
+//cmd ::= SPLIT VGROUP NK_INTEGER(A).                                               { pCxt->pRootNode = createSplitVgroupStmt(pCxt, &A); }
 
 %type dnode_list                                                                  { SNodeList* }
 %destructor dnode_list                                                            { nodesDestroyList($$); }

source/libs/parser/src/parAstCreater.c

@@ -1435,25 +1435,37 @@ SNode* createKillStmt(SAstCreateContext* pCxt, ENodeType type, const SToken* pId
   return (SNode*)pStmt;
 }
 
+SNode* createBalanceVgroupStmt(SAstCreateContext* pCxt) {
+  CHECK_PARSER_STATUS(pCxt);
+  SBalanceVgroupStmt* pStmt = nodesMakeNode(QUERY_NODE_BALANCE_VGROUP_STMT);
+  CHECK_OUT_OF_MEM(pStmt);
+  return (SNode*)pStmt;
+}
+
 SNode* createMergeVgroupStmt(SAstCreateContext* pCxt, const SToken* pVgId1, const SToken* pVgId2) {
   CHECK_PARSER_STATUS(pCxt);
-  SNode* pStmt = nodesMakeNode(QUERY_NODE_MERGE_VGROUP_STMT);
+  SMergeVgroupStmt* pStmt = nodesMakeNode(QUERY_NODE_MERGE_VGROUP_STMT);
   CHECK_OUT_OF_MEM(pStmt);
-  return pStmt;
+  pStmt->vgId1 = taosStr2Int32(pVgId1->z, NULL, 10);
+  pStmt->vgId2 = taosStr2Int32(pVgId2->z, NULL, 10);
+  return (SNode*)pStmt;
 }
 
 SNode* createRedistributeVgroupStmt(SAstCreateContext* pCxt, const SToken* pVgId, SNodeList* pDnodes) {
   CHECK_PARSER_STATUS(pCxt);
-  SNode* pStmt = nodesMakeNode(QUERY_NODE_REDISTRIBUTE_VGROUP_STMT);
+  SRedistributeVgroupStmt* pStmt = nodesMakeNode(QUERY_NODE_REDISTRIBUTE_VGROUP_STMT);
   CHECK_OUT_OF_MEM(pStmt);
-  return pStmt;
+  pStmt->vgId = taosStr2Int32(pVgId->z, NULL, 10);
+  pStmt->pDnodes = pDnodes;
+  return (SNode*)pStmt;
 }
 
 SNode* createSplitVgroupStmt(SAstCreateContext* pCxt, const SToken* pVgId) {
   CHECK_PARSER_STATUS(pCxt);
-  SNode* pStmt = nodesMakeNode(QUERY_NODE_SPLIT_VGROUP_STMT);
+  SSplitVgroupStmt* pStmt = nodesMakeNode(QUERY_NODE_SPLIT_VGROUP_STMT);
   CHECK_OUT_OF_MEM(pStmt);
-  return pStmt;
+  pStmt->vgId = taosStr2Int32(pVgId->z, NULL, 10);
+  return (SNode*)pStmt;
 }
 
 SNode* createSyncdbStmt(SAstCreateContext* pCxt, const SToken* pDbName) {

source/libs/parser/src/parTokenizer.c

@@ -41,6 +41,7 @@ static SKeyword keywordTable[] = {
     {"AS",            TK_AS},
     {"ASC",           TK_ASC},
     {"AT_ONCE",       TK_AT_ONCE},
+    {"BALANCE",       TK_BALANCE},
     {"BETWEEN",       TK_BETWEEN},
     {"BINARY",        TK_BINARY},
     {"BIGINT",        TK_BIGINT},
@@ -116,6 +117,7 @@ static SKeyword keywordTable[] = {
     {"LOCAL",         TK_LOCAL},
     {"MATCH",         TK_MATCH},
     {"MAXROWS",       TK_MAXROWS},
+    {"MERGE",         TK_MERGE},
     {"MINROWS",       TK_MINROWS},
     {"MINUS",         TK_MINUS},
     {"MNODE",         TK_MNODE},
@@ -151,6 +153,7 @@ static SKeyword keywordTable[] = {
     {"QUERY",         TK_QUERY},
     {"RATIO",         TK_RATIO},
     {"READ",          TK_READ},
+    {"REDISTRIBUTE",  TK_REDISTRIBUTE},
     {"RENAME",        TK_RENAME},
     {"REPLICA",       TK_REPLICA},
     {"RESET",         TK_RESET},
@@ -171,6 +174,7 @@ static SKeyword keywordTable[] = {
     {"SNODE",         TK_SNODE},
     {"SNODES",        TK_SNODES},
     {"SOFFSET",       TK_SOFFSET},
+    // {"SPLIT",         TK_SPLIT},
     {"STABLE",        TK_STABLE},
     {"STABLES",       TK_STABLES},
     {"STATE",         TK_STATE},
@@ -208,6 +212,7 @@ static SKeyword keywordTable[] = {
     {"VARCHAR",       TK_VARCHAR},
     {"VARIABLES",     TK_VARIABLES},
     {"VERBOSE",       TK_VERBOSE},
+    {"VGROUP",        TK_VGROUP},
     {"VGROUPS",       TK_VGROUPS},
     {"VNODES",        TK_VNODES},
     {"WAL",           TK_WAL},

source/libs/parser/src/parTranslater.c

@@ -847,9 +847,9 @@ static EDealRes translateJsonOperator(STranslateContext* pCxt, SOperatorNode* pO
   if (TSDB_DATA_TYPE_JSON != ldt.type || TSDB_DATA_TYPE_BINARY != rdt.type) {
     return generateDealNodeErrMsg(pCxt, TSDB_CODE_PAR_WRONG_VALUE_TYPE, ((SExprNode*)(pOp->pRight))->aliasName);
   }
-  if(pOp->opType == OP_TYPE_JSON_GET_VALUE){
+  if (pOp->opType == OP_TYPE_JSON_GET_VALUE) {
     pOp->node.resType.type = TSDB_DATA_TYPE_JSON;
-  }else if(pOp->opType == OP_TYPE_JSON_CONTAINS){
+  } else if (pOp->opType == OP_TYPE_JSON_CONTAINS) {
     pOp->node.resType.type = TSDB_DATA_TYPE_BOOL;
   }
   pOp->node.resType.bytes = tDataTypes[pOp->node.resType.type].bytes;
@@ -3611,6 +3611,55 @@ static int32_t translateRevoke(STranslateContext* pCxt, SRevokeStmt* pStmt) {
   return buildCmdMsg(pCxt, TDMT_MND_ALTER_USER, (FSerializeFunc)tSerializeSAlterUserReq, &req);
 }
 
+static int32_t translateBalanceVgroup(STranslateContext* pCxt, SBalanceVgroupStmt* pStmt) {
+  SBalanceVgroupReq req = {0};
+  return buildCmdMsg(pCxt, TDMT_MND_BALANCE_VGROUP, (FSerializeFunc)tSerializeSBalanceVgroupReq, &req);
+}
+
+static int32_t translateMergeVgroup(STranslateContext* pCxt, SMergeVgroupStmt* pStmt) {
+  SMergeVgroupReq req = {.vgId1 = pStmt->vgId1, .vgId2 = pStmt->vgId2};
+  return buildCmdMsg(pCxt, TDMT_MND_MERGE_VGROUP, (FSerializeFunc)tSerializeSMergeVgroupReq, &req);
+}
+
+static int32_t checkDnodeIds(STranslateContext* pCxt, SRedistributeVgroupStmt* pStmt) {
+  int32_t numOfDnodes = LIST_LENGTH(pStmt->pDnodes);
+  if (numOfDnodes > 3 || numOfDnodes < 1) {
+    return generateSyntaxErrMsg(&pCxt->msgBuf, TSDB_CODE_PAR_INVALID_REDISTRIBUTE_VG);
+  }
+
+  SNode* pNode = NULL;
+  FOREACH(pNode, pStmt->pDnodes) {
+    SValueNode* pVal = (SValueNode*)pNode;
+    if (DEAL_RES_ERROR == translateValue(pCxt, pVal)) {
+      return pCxt->errCode;
+    }
+  }
+
+  pStmt->dnodeId1 = getBigintFromValueNode((SValueNode*)nodesListGetNode(pStmt->pDnodes, 0));
+  pStmt->dnodeId2 = -1;
+  pStmt->dnodeId3 = -1;
+  if (numOfDnodes > 1) {
+    pStmt->dnodeId2 = getBigintFromValueNode((SValueNode*)nodesListGetNode(pStmt->pDnodes, 1));
+  }
+  if (numOfDnodes > 2) {
+    pStmt->dnodeId3 = getBigintFromValueNode((SValueNode*)nodesListGetNode(pStmt->pDnodes, 2));
+  }
+
+  return TSDB_CODE_SUCCESS;
+}
+
+static int32_t translateRedistributeVgroup(STranslateContext* pCxt, SRedistributeVgroupStmt* pStmt) {
+  SRedistributeVgroupReq req = {.vgId = pStmt->vgId};
+  int32_t                code = checkDnodeIds(pCxt, pStmt);
+  if (TSDB_CODE_SUCCESS == code) {
+    req.dnodeId1 = pStmt->dnodeId1;
+    req.dnodeId2 = pStmt->dnodeId2;
+    req.dnodeId3 = pStmt->dnodeId3;
+    code = buildCmdMsg(pCxt, TDMT_MND_REDISTRIBUTE_VGROUP, (FSerializeFunc)tSerializeSRedistributeVgroupReq, &req);
+  }
+  return code;
+}
+
 static int32_t translateQuery(STranslateContext* pCxt, SNode* pNode) {
   int32_t code = TSDB_CODE_SUCCESS;
   switch (nodeType(pNode)) {
@@ -3733,6 +3782,15 @@ static int32_t translateQuery(STranslateContext* pCxt, SNode* pNode) {
     case QUERY_NODE_REVOKE_STMT:
       code = translateRevoke(pCxt, (SRevokeStmt*)pNode);
       break;
+    case QUERY_NODE_BALANCE_VGROUP_STMT:
+      code = translateBalanceVgroup(pCxt, (SBalanceVgroupStmt*)pNode);
+      break;
+    case QUERY_NODE_MERGE_VGROUP_STMT:
+      code = translateMergeVgroup(pCxt, (SMergeVgroupStmt*)pNode);
+      break;
+    case QUERY_NODE_REDISTRIBUTE_VGROUP_STMT:
+      code = translateRedistributeVgroup(pCxt, (SRedistributeVgroupStmt*)pNode);
+      break;
     default:
       break;
   }

source/libs/parser/src/parUtil.c

@@ -180,6 +180,8 @@ static char* getSyntaxErrFormat(int32_t errCode) {
       return "Value too long for column/tag: %s";
     case TSDB_CODE_PAR_INVALID_DELETE_WHERE:
       return "The DELETE statement must have a definite time window range";
+    case TSDB_CODE_PAR_INVALID_REDISTRIBUTE_VG:
+      return "The REDISTRIBUTE VGROUP statement only support 1 to 3 dnodes";
     case TSDB_CODE_OUT_OF_MEMORY:
       return "Out of memory";
     default:

source/libs/parser/test/parExplainToSyncdbTest.cpp

@@ -19,7 +19,7 @@ using namespace std;
 
 namespace ParserTest {
 
-class ParserExplainToSyncdbTest : public ParserTestBase {};
+class ParserExplainToSyncdbTest : public ParserDdlTest {};
 
 TEST_F(ParserExplainToSyncdbTest, explain) {
   useDb("root", "test");
@@ -43,8 +43,62 @@ TEST_F(ParserExplainToSyncdbTest, grant) {
 // todo kill connection
 // todo kill query
 // todo kill stream
-// todo merge vgroup
-// todo redistribute vgroup
+
+TEST_F(ParserExplainToSyncdbTest, mergeVgroup) {
+  useDb("root", "test");
+
+  SMergeVgroupReq expect = {0};
+
+  auto setMergeVgroupReqFunc = [&](int32_t vgId1, int32_t vgId2) {
+    expect.vgId1 = vgId1;
+    expect.vgId2 = vgId2;
+  };
+
+  setCheckDdlFunc([&](const SQuery* pQuery, ParserStage stage) {
+    ASSERT_EQ(nodeType(pQuery->pRoot), QUERY_NODE_MERGE_VGROUP_STMT);
+    ASSERT_EQ(pQuery->pCmdMsg->msgType, TDMT_MND_MERGE_VGROUP);
+    SMergeVgroupReq req = {0};
+    ASSERT_EQ(tDeserializeSMergeVgroupReq(pQuery->pCmdMsg->pMsg, pQuery->pCmdMsg->msgLen, &req), TSDB_CODE_SUCCESS);
+    ASSERT_EQ(req.vgId1, expect.vgId1);
+    ASSERT_EQ(req.vgId2, expect.vgId2);
+  });
+
+  setMergeVgroupReqFunc(1, 2);
+  run("MERGE VGROUP 1 2");
+}
+
+TEST_F(ParserExplainToSyncdbTest, redistributeVgroup) {
+  useDb("root", "test");
+
+  SRedistributeVgroupReq expect = {0};
+
+  auto setRedistributeVgroupReqFunc = [&](int32_t vgId, int32_t dnodeId1, int32_t dnodeId2 = -1,
+                                          int32_t dnodeId3 = -1) {
+    expect.vgId = vgId;
+    expect.dnodeId1 = dnodeId1;
+    expect.dnodeId2 = dnodeId2;
+    expect.dnodeId3 = dnodeId3;
+  };
+
+  setCheckDdlFunc([&](const SQuery* pQuery, ParserStage stage) {
+    ASSERT_EQ(nodeType(pQuery->pRoot), QUERY_NODE_REDISTRIBUTE_VGROUP_STMT);
+    ASSERT_EQ(pQuery->pCmdMsg->msgType, TDMT_MND_REDISTRIBUTE_VGROUP);
+    SRedistributeVgroupReq req = {0};
+    ASSERT_EQ(tDeserializeSRedistributeVgroupReq(pQuery->pCmdMsg->pMsg, pQuery->pCmdMsg->msgLen, &req),
+              TSDB_CODE_SUCCESS);
+    ASSERT_EQ(req.vgId, expect.vgId);
+    ASSERT_EQ(req.dnodeId1, expect.dnodeId1);
+    ASSERT_EQ(req.dnodeId2, expect.dnodeId2);
+    ASSERT_EQ(req.dnodeId3, expect.dnodeId3);
+  });
+
+  setRedistributeVgroupReqFunc(3, 1);
+  run("REDISTRIBUTE VGROUP 3 DNODE 1");
+
+  setRedistributeVgroupReqFunc(5, 10, 20, 30);
+  run("REDISTRIBUTE VGROUP 5 DNODE 10 DNODE 20 DNODE 30");
+}
+
 // todo reset query cache
 
 TEST_F(ParserExplainToSyncdbTest, revoke) {

source/libs/parser/test/parInitialATest.cpp

@@ -305,6 +305,19 @@ TEST_F(ParserInitialATest, alterUser) {
   run("ALTER user wxy privilege 'write'");
 }
 
+TEST_F(ParserInitialATest, balanceVgroup) {
+  useDb("root", "test");
+
+  setCheckDdlFunc([&](const SQuery* pQuery, ParserStage stage) {
+    ASSERT_EQ(nodeType(pQuery->pRoot), QUERY_NODE_BALANCE_VGROUP_STMT);
+    ASSERT_EQ(pQuery->pCmdMsg->msgType, TDMT_MND_BALANCE_VGROUP);
+    SBalanceVgroupReq req = {0};
+    ASSERT_EQ(tDeserializeSBalanceVgroupReq(pQuery->pCmdMsg->pMsg, pQuery->pCmdMsg->msgLen, &req), TSDB_CODE_SUCCESS);
+  });
+
+  run("BALANCE VGROUP");
+}
+
 TEST_F(ParserInitialATest, bug001) {
   useDb("root", "test");
 

source/libs/stream/src/streamData.c

@@ -13,7 +13,7 @@
  * along with this program. If not, see <http://www.gnu.org/licenses/>.
  */
 
-#include "tstream.h"
+#include "streamInc.h"
 
 #if 0
 int32_t streamDataBlockEncode(void** buf, const SStreamDataBlock* pOutput) {
@@ -74,7 +74,6 @@ FAIL:
 }
 
 static FORCE_INLINE void streamDataSubmitRefInc(SStreamDataSubmit* pDataSubmit) {
-  //
   atomic_add_fetch_32(pDataSubmit->dataRef, 1);
 }
 

source/libs/stream/src/streamMsg.c

@@ -116,11 +116,9 @@ int32_t streamBuildDispatchMsg(SStreamTask* pTask, SStreamDataBlock* data, SRpcM
     *ppEpSet = &pTask->fixedEpDispatcher.epSet;
     downstreamTaskId = pTask->fixedEpDispatcher.taskId;
   } else if (pTask->dispatchType == TASK_DISPATCH__SHUFFLE) {
-    // TODO get ctbName
-    char         ctbName[TSDB_TABLE_FNAME_LEN + 22] = {0};
+    // TODO get ctbName for each block
     SSDataBlock* pBlock = taosArrayGet(data->blocks, 0);
-    sprintf(ctbName, "%s:%ld", pTask->shuffleDispatcher.stbFullName, pBlock->info.groupId);
-    // get vg and ep
+    char*        ctbName = buildCtbNameByGroupId(pTask->shuffleDispatcher.stbFullName, pBlock->info.groupId);
     // TODO: get hash function by hashMethod
 
     // get groupId, compute hash value

source/util/src/terror.c

@@ -451,9 +451,89 @@ TAOS_DEFINE_ERROR(TSDB_CODE_SCH_TIMEOUT_ERROR,            "Task timeout")
 TAOS_DEFINE_ERROR(TSDB_CODE_QW_MSG_ERROR,                 "Invalid msg order")
 
 // parser
-TAOS_DEFINE_ERROR(TSDB_CODE_PAR_TABLE_NOT_EXIST,          "Table does not exist")
-TAOS_DEFINE_ERROR(TSDB_CODE_PAR_PERMISSION_DENIED,        "Permission denied")
-TAOS_DEFINE_ERROR(TSDB_CODE_PAR_INTERNAL_ERROR,           "Parser internal error")
+TAOS_DEFINE_ERROR(TSDB_CODE_PAR_PERMISSION_DENIED,         "Permission denied")
+TAOS_DEFINE_ERROR(TSDB_CODE_PAR_INTERNAL_ERROR,            "Parser internal error")
+TAOS_DEFINE_ERROR(TSDB_CODE_PAR_SYNTAX_ERROR,              "syntax error near")
+TAOS_DEFINE_ERROR(TSDB_CODE_PAR_INCOMPLETE_SQL,            "Incomplete SQL statement")
+TAOS_DEFINE_ERROR(TSDB_CODE_PAR_INVALID_COLUMN,            "Invalid column name")
+TAOS_DEFINE_ERROR(TSDB_CODE_PAR_TABLE_NOT_EXIST,           "Table does not exist")
+TAOS_DEFINE_ERROR(TSDB_CODE_PAR_AMBIGUOUS_COLUMN,          "Column ambiguously defined")
+TAOS_DEFINE_ERROR(TSDB_CODE_PAR_WRONG_VALUE_TYPE,          "Invalid value type")
+TAOS_DEFINE_ERROR(TSDB_CODE_PAR_ILLEGAL_USE_AGG_FUNCTION,  "There mustn't be aggregation")
+TAOS_DEFINE_ERROR(TSDB_CODE_PAR_WRONG_NUMBER_OF_SELECT,    "ORDER BY item must be the number of a SELECT-list expression")
+TAOS_DEFINE_ERROR(TSDB_CODE_PAR_GROUPBY_LACK_EXPRESSION,   "Not a GROUP BY expression")
+TAOS_DEFINE_ERROR(TSDB_CODE_PAR_NOT_SELECTED_EXPRESSION,   "Not SELECTed expression")
+TAOS_DEFINE_ERROR(TSDB_CODE_PAR_NOT_SINGLE_GROUP,          "Not a single-group group function")
+TAOS_DEFINE_ERROR(TSDB_CODE_PAR_TAGS_NOT_MATCHED,          "Tags number not matched")
+TAOS_DEFINE_ERROR(TSDB_CODE_PAR_INVALID_TAG_NAME,          "Invalid tag name")
+TAOS_DEFINE_ERROR(TSDB_CODE_PAR_NAME_OR_PASSWD_TOO_LONG,   "Name or password too long")
+TAOS_DEFINE_ERROR(TSDB_CODE_PAR_PASSWD_EMPTY,              "Password can not be empty")
+TAOS_DEFINE_ERROR(TSDB_CODE_PAR_INVALID_PORT,              "Port should be an integer that is less than 65535 and greater than 0")
+TAOS_DEFINE_ERROR(TSDB_CODE_PAR_INVALID_ENDPOINT,          "Endpoint should be in the format of 'fqdn:port'")
+TAOS_DEFINE_ERROR(TSDB_CODE_PAR_EXPRIE_STATEMENT,          "This statement is no longer supported")
+TAOS_DEFINE_ERROR(TSDB_CODE_PAR_INTER_VALUE_TOO_SMALL,     "Interval too small")
+TAOS_DEFINE_ERROR(TSDB_CODE_PAR_DB_NOT_SPECIFIED,          "Database not specified")
+TAOS_DEFINE_ERROR(TSDB_CODE_PAR_INVALID_IDENTIFIER_NAME,   "Invalid identifier name")
+TAOS_DEFINE_ERROR(TSDB_CODE_PAR_CORRESPONDING_STABLE_ERR,  "Corresponding super table not in this db")
+TAOS_DEFINE_ERROR(TSDB_CODE_PAR_INVALID_RANGE_OPTION,      "Invalid option")
+TAOS_DEFINE_ERROR(TSDB_CODE_PAR_INVALID_STR_OPTION,        "Invalid option")
+TAOS_DEFINE_ERROR(TSDB_CODE_PAR_INVALID_ENUM_OPTION,       "Invalid option")
+TAOS_DEFINE_ERROR(TSDB_CODE_PAR_INVALID_KEEP_NUM,          "Invalid number of keep options")
+TAOS_DEFINE_ERROR(TSDB_CODE_PAR_INVALID_KEEP_ORDER,        "Invalid keep value, should be keep0 <= keep1 <= keep2")
+TAOS_DEFINE_ERROR(TSDB_CODE_PAR_INVALID_KEEP_VALUE,        "Invalid option keep")
+TAOS_DEFINE_ERROR(TSDB_CODE_PAR_INVALID_COMMENT_OPTION,    "Invalid option comment")
+TAOS_DEFINE_ERROR(TSDB_CODE_PAR_INVALID_F_RANGE_OPTION,    "Invalid option")
+TAOS_DEFINE_ERROR(TSDB_CODE_PAR_INVALID_ROLLUP_OPTION,     "Invalid option rollup: only one function is allowed")
+TAOS_DEFINE_ERROR(TSDB_CODE_PAR_INVALID_RETENTIONS_OPTION, "Invalid option retentions")
+TAOS_DEFINE_ERROR(TSDB_CODE_PAR_GROUPBY_WINDOW_COEXIST,    "GROUP BY and WINDOW-clause can't be used together")
+TAOS_DEFINE_ERROR(TSDB_CODE_PAR_INVALID_OPTION_UNIT,       "Invalid option unit: only m, h, d allowed")
+TAOS_DEFINE_ERROR(TSDB_CODE_PAR_INVALID_KEEP_UNIT,         "Invalid option keep unit: only m, h, d allowed")
+TAOS_DEFINE_ERROR(TSDB_CODE_PAR_AGG_FUNC_NESTING,          "Aggregate functions do not support nesting")
+TAOS_DEFINE_ERROR(TSDB_CODE_PAR_INVALID_STATE_WIN_TYPE,    "Only support STATE_WINDOW on integer/bool/varchar column")
+TAOS_DEFINE_ERROR(TSDB_CODE_PAR_INVALID_STATE_WIN_COL,     "Not support STATE_WINDOW on tag column")
+TAOS_DEFINE_ERROR(TSDB_CODE_PAR_INVALID_STATE_WIN_TABLE,   "STATE_WINDOW not support for super table query")
+TAOS_DEFINE_ERROR(TSDB_CODE_PAR_INTER_SESSION_GAP,         "SESSION gap should be fixed time window, and greater than 0")
+TAOS_DEFINE_ERROR(TSDB_CODE_PAR_INTER_SESSION_COL,         "Only support SESSION on primary timestamp column")
+TAOS_DEFINE_ERROR(TSDB_CODE_PAR_INTER_OFFSET_NEGATIVE,     "Interval offset cannot be negative")
+TAOS_DEFINE_ERROR(TSDB_CODE_PAR_INTER_OFFSET_UNIT,         "Cannot use 'year' as offset when interval is 'month'")
+TAOS_DEFINE_ERROR(TSDB_CODE_PAR_INTER_OFFSET_TOO_BIG,      "Interval offset should be shorter than interval")
+TAOS_DEFINE_ERROR(TSDB_CODE_PAR_INTER_SLIDING_UNIT,        "Does not support sliding when interval is natural month/year")
+TAOS_DEFINE_ERROR(TSDB_CODE_PAR_INTER_SLIDING_TOO_BIG,     "sliding value no larger than the interval value")
+TAOS_DEFINE_ERROR(TSDB_CODE_PAR_INTER_SLIDING_TOO_SMALL,   "sliding value can not less than 1% of interval value")
+TAOS_DEFINE_ERROR(TSDB_CODE_PAR_ONLY_ONE_JSON_TAG,         "Only one tag if there is a json tag")
+TAOS_DEFINE_ERROR(TSDB_CODE_PAR_INCORRECT_NUM_OF_COL,      "Query block has incorrect number of result columns")
+TAOS_DEFINE_ERROR(TSDB_CODE_PAR_INCORRECT_TIMESTAMP_VAL,   "Incorrect TIMESTAMP value")
+TAOS_DEFINE_ERROR(TSDB_CODE_PAR_INVALID_DAYS_VALUE,        "Invalid days value, should be keep2 >= keep1 >= keep0 >= days")
+TAOS_DEFINE_ERROR(TSDB_CODE_PAR_OFFSET_LESS_ZERO,          "soffset/offset can not be less than 0")
+TAOS_DEFINE_ERROR(TSDB_CODE_PAR_SLIMIT_LEAK_PARTITION_BY,  "slimit/soffset only available for PARTITION BY query")
+TAOS_DEFINE_ERROR(TSDB_CODE_PAR_INVALID_TOPIC_QUERY,        "Invalid topic query")
+TAOS_DEFINE_ERROR(TSDB_CODE_PAR_INVALID_DROP_STABLE,        "Cannot drop super table in batch")
+TAOS_DEFINE_ERROR(TSDB_CODE_PAR_INVALID_FILL_TIME_RANGE,    "Start(end) time of query range required or time range too large")
+TAOS_DEFINE_ERROR(TSDB_CODE_PAR_DUPLICATED_COLUMN,          "Duplicated column names")
+TAOS_DEFINE_ERROR(TSDB_CODE_PAR_INVALID_TAGS_LENGTH,        "Tags length exceeds max length")
+TAOS_DEFINE_ERROR(TSDB_CODE_PAR_INVALID_ROW_LENGTH,         "Row length exceeds max length")
+TAOS_DEFINE_ERROR(TSDB_CODE_PAR_INVALID_COLUMNS_NUM,        "Illegal number of columns")
+TAOS_DEFINE_ERROR(TSDB_CODE_PAR_TOO_MANY_COLUMNS,           "Too many columns")
+TAOS_DEFINE_ERROR(TSDB_CODE_PAR_INVALID_FIRST_COLUMN,       "First column must be timestamp")
+TAOS_DEFINE_ERROR(TSDB_CODE_PAR_INVALID_VAR_COLUMN_LEN,     "Invalid binary/nchar column length")
+TAOS_DEFINE_ERROR(TSDB_CODE_PAR_INVALID_TAGS_NUM,           "Invalid number of tag columns")
+TAOS_DEFINE_ERROR(TSDB_CODE_PAR_INVALID_INTERNAL_PK,        "Invalid _c0 or _rowts expression")
+TAOS_DEFINE_ERROR(TSDB_CODE_PAR_INVALID_TIMELINE_FUNC,      "Invalid timeline function")
+TAOS_DEFINE_ERROR(TSDB_CODE_PAR_INVALID_PASSWD,             "Invalid password")
+TAOS_DEFINE_ERROR(TSDB_CODE_PAR_INVALID_ALTER_TABLE,        "Invalid alter table statement")
+TAOS_DEFINE_ERROR(TSDB_CODE_PAR_CANNOT_DROP_PRIMARY_KEY,    "Primary timestamp column cannot be dropped")
+TAOS_DEFINE_ERROR(TSDB_CODE_PAR_INVALID_MODIFY_COL,         "Only binary/nchar column length could be modified")
+TAOS_DEFINE_ERROR(TSDB_CODE_PAR_INVALID_TBNAME,             "Invalid tbname pseudo column")
+TAOS_DEFINE_ERROR(TSDB_CODE_PAR_INVALID_FUNCTION_NAME,      "Invalid function name")
+TAOS_DEFINE_ERROR(TSDB_CODE_PAR_COMMENT_TOO_LONG,           "Comment too long")
+TAOS_DEFINE_ERROR(TSDB_CODE_PAR_NOT_ALLOWED_FUNC,           "Some functions are allowed only in the SELECT list of a query. "
+                                                            "And, cannot be mixed with other non scalar functions or columns.")
+TAOS_DEFINE_ERROR(TSDB_CODE_PAR_NOT_ALLOWED_WIN_QUERY,      "Window query not supported, since the result of subquery not include valid timestamp column")
+TAOS_DEFINE_ERROR(TSDB_CODE_PAR_INVALID_DROP_COL,           "No columns can be dropped")
+TAOS_DEFINE_ERROR(TSDB_CODE_PAR_INVALID_COL_JSON,           "Only tag can be json type")
+TAOS_DEFINE_ERROR(TSDB_CODE_PAR_VALUE_TOO_LONG,             "Value too long for column/tag")
+TAOS_DEFINE_ERROR(TSDB_CODE_PAR_INVALID_DELETE_WHERE,       "The DELETE statement must have a definite time window range")
+TAOS_DEFINE_ERROR(TSDB_CODE_PAR_INVALID_REDISTRIBUTE_VG,    "The REDISTRIBUTE VGROUP statement only support 1 to 3 dnodes")
 
 //planner
 TAOS_DEFINE_ERROR(TSDB_CODE_PLAN_INTERNAL_ERROR,          "Planner internal error")

tests/script/tsim/sync/oneReplica1VgElectWithInsert.sim

@@ -31,7 +31,7 @@ if $data[0][4] != ready then
   goto check_dnode_ready
 endi
 
-#sql connect
+sql connect
 sql create dnode $hostname port 7200
 sql create dnode $hostname port 7300
 sql create dnode $hostname port 7400
@@ -71,7 +71,7 @@ sql create database db replica $replica vgroups $vgroups
 $loop_cnt = 0
 check_db_ready:
 $loop_cnt = $loop_cnt + 1
-sleep 200
+sleep 20
 if $loop_cnt == 10 then
   print ====> db not ready!
 	return -1
@@ -93,13 +93,12 @@ $loop_cnt = 0
 check_vg_ready:
 $loop_cnt = $loop_cnt + 1
 sleep 200
-if $loop_cnt == 10 then
+if $loop_cnt == 300 then
   print ====> vgroups not ready!
 	return -1
 endi
 sql show vgroups
 print ===> rows: $rows
-print $data(2)[0] $data(2)[1] $data(2)[2] $data(2)[3] $data(2)[4] $data(2)[5] $data(2)[6] $data(2)[7] $data(2)[8] $data(2)[9] $data(2)[10] $data(2)[11] $data(2)[12] $data(2)[13]
 print $data[0][0] $data[0][1] $data[0][2] $data[0][3] $data[0][4] $data[0][5] $data[0][6] $data[0][7] $data[0][8] $data[0][9] $data[10][6] $data[0][11] $data[0][12] $data[0][13]
 if $rows != $vgroups then
   return -1
@@ -132,10 +131,12 @@ if $data[0][8] == leader then
      goto check_vg_ready
    endi    
    print ---- vgroup $data[0][0] leader locate on dnode $data[0][7] 
-  goto vg_ready
+   goto vg_ready
+else
+   goto check_vg_ready
 endi
-vg_ready:
 
+vg_ready:
 print ====>  create stable/child table
 sql create table stb (ts timestamp, c1 int, c2 float, c3 binary(10)) tags (t1 int)
 
@@ -156,27 +157,13 @@ while $i < $tbNum
   sql create table $ctb using stb tags( $i )
   $ntb = $ntbPrefix . $i
   sql create table $ntb (ts timestamp, c1 int, c2 float, c3 binary(10))
-  
-#  $x = 0
-#  while $x < $rowNum
-#    $binary = ' . binary
-#    $binary = $binary . $i
-#    $binary = $binary . '
-#
-#    sql insert into $ctb values ($tstart , $i , $x , $binary )
-#    sql insert into $ntb values ($tstart , 999 , 999 , 'binary-ntb' )
-#    $tstart = $tstart + 1
-#    $x = $x + 1
-#  endw
-  
-#  print ====> insert rows: $rowNum into $ctb and $ntb
-
   $i = $i + 1
-#  $tstart = 1640966400000
 endw
 
 $totalTblNum = $tbNum * 2
+sleep 1000
 sql show tables
+print ====> expect $totalTblNum and infinsert $rows  in fact 
 if $rows != $totalTblNum then 
   return -1
 endi
@@ -185,7 +172,7 @@ print ====>  create a normal table for interaction between main and back threads
 sql create table interaction (ts timestamp, flag binary(10), childrows int, stbrows int)
 
 print ====> start to run_back to insert data
-run_back tsim/tmq/insertDataByRunBack.sim
+run_back tsim/sync/insertDataByRunBack.sim
 
 
 print ====> waiting insert thread starting insert data
@@ -222,12 +209,15 @@ endi
 $dnodeId = dnode . $dnodeId
 print ====> stop $dnodeId
 system sh/exec.sh -n $dnodeId -s stop -x SIGINT
+sleep 1000
+print ====> start $dnodeId
+system sh/exec.sh -n $dnodeId -s start
 
 $loop_cnt = 0
 check_vg_ready_2:
 $loop_cnt = $loop_cnt + 1
 sleep 200
-if $loop_cnt == 10 then
+if $loop_cnt == 300 then
   print ====> vgroups switch fail!!!
 	return -1
 endi
@@ -242,38 +232,38 @@ if $data[0][4] == leader then
    if $data[0][6] != NULL then
      goto check_vg_ready_2
    endi
-   if $data[0][8] != NULL then
+   if $data[0][8] != FOLLOWER then
      goto check_vg_ready_2
    endi
-   print ---- vgroup $data[0][0] leader switch to dnode $data[0][3] 
+   print ---- vgroup $dnodeId leader switch to dnode $data[0][3] 
   goto vg_ready_2
 endi
 if $data[0][6] == leader then
    if $data[0][4] != NULL then
      goto check_vg_ready_2
    endi
-   if $data[0][8] != NULL then
+   if $data[0][8] != FOLLOWER then
      goto check_vg_ready_2
    endi    
-   print ---- vgroup $data[0][0] leader switch to dnode $data[0][5] 
+   print ---- vgroup $dnodeId leader switch to dnode $data[0][5] 
   goto vg_ready_2
 endi
 if $data[0][8] == leader then
    if $data[0][4] != NULL then
      goto check_vg_ready_2
    endi
-   if $data[0][6] != NULL then
+   if $data[0][6] != FOLLOWER then
      goto check_vg_ready_2
    endi    
-   print ---- vgroup $data[0][0] leader switch to dnode $data[0][7] 
+   print ---- vgroup $dnodeId leader switch to dnode $data[0][7] 
   goto vg_ready_2
+else 
+  goto check_vg_ready_2
 endi
 vg_ready_2:
 
 $switch_loop_cnt = $switch_loop_cnt + 1
 if $switch_loop_cnt < 3 then
-  print ====> start $dnodeId
-  system sh/exec.sh -n $dnodeId -s start
   goto switch_leader_loop
 endi
 
@@ -347,7 +337,7 @@ if $data[0][4] == leader then
    if $data[0][6] != NULL then
      goto check_vg_ready_1
    endi
-   if $data[0][8] != NULL then
+   if $data[0][8] != FOLLOWER then
      goto check_vg_ready_1
    endi    
   goto vg_ready_1
@@ -356,7 +346,7 @@ if $data[0][6] == leader then
    if $data[0][4] != NULL then
      goto check_vg_ready_1
    endi
-   if $data[0][8] != NULL then
+   if $data[0][8] != FOLLOWER then
      goto check_vg_ready_1
    endi    
   goto vg_ready_1
@@ -365,7 +355,7 @@ if $data[0][8] == leader then
    if $data[0][4] != NULL then
      goto check_vg_ready_1
    endi
-   if $data[0][6] != NULL then
+   if $data[0][6] != FOLLOWER then
      goto check_vg_ready_1
    endi    
   goto vg_ready_1
@@ -394,6 +384,73 @@ if $data[0][0] != $totalRowsOfStb then
   return -1
 endi
 
+print ====> once stop one dnode by loop, and do query every time
+$i = 2
+loop_stop_dnode:
+$dnodeId = dnode . $i
+print ====> stop $dnodeId
+system sh/exec.sh -n $dnodeId -s stop -x SIGINT
+
+check_vg_ready_3:
+sql show vgroups
+print $data(2)[0] $data(2)[1] $data(2)[2] $data(2)[3] $data(2)[4] $data(2)[5] $data(2)[6] $data(2)[7] $data(2)[8] $data(2)[9] $data(2)[10] $data(2)[11] $data(2)[12] $data(2)[13]
+if $data[0][4] == LEADER then
+  if $data[0][6] == LEADER then
+    goto check_vg_ready_3
+  endi
+  if $data[0][8] == LEADER then
+    goto check_vg_ready_3
+  endi
+  print ---- vgroup $data[0][0] leader locating dnode $data[0][5]
+elif $data[0][6] == LEADER then
+  if $data[0][4] == LEADER then
+    goto check_vg_ready_3
+  endi
+  if $data[0][8] == LEADER then
+    goto check_vg_ready_3
+  endi
+  print ---- vgroup $data[0][0] leader locating dnode $data[0][7]
+elif $data[0][8] == LEADER then
+  if $data[0][4] == LEADER then
+    goto check_vg_ready_3
+  endi
+  if $data[0][6] == LEADER then
+    goto check_vg_ready_3
+  endi
+  print ---- vgroup $data[0][0] leader locating dnode $data[0][9]
+else 
+   print ====> no leader vnode!!!
+  return -1 
+endi
+
+sql select count(*) from ntb0
+print rows: $rows 
+print $data[0][0]  $data[0][1]
+if $data[0][0] != $totalRowsOfCtb then 
+  return -1
+endi
+
+sql select count(*) from ctb0
+print rows: $rows 
+print $data[0][0]  $data[0][1]
+if $data[0][0] != $totalRowsOfCtb then 
+  return -1
+endi
+
+sql select count(*) from stb
+print rows: $rows 
+print $data[0][0]  $data[0][1]
+if $data[0][0] != $totalRowsOfStb then 
+  return -1
+endi
+
+$i = $i + 1
+if $i <= 4 then 
+  print ====> start $dnodeId
+  system sh/exec.sh -n $dnodeId -s start
+  goto loop_stop_dnode
+endi
+
 system sh/exec.sh -n dnode1 -s stop -x SIGINT
 system sh/exec.sh -n dnode2 -s stop -x SIGINT
 system sh/exec.sh -n dnode3 -s stop -x SIGINT

tests/system-test/1-insert/test_stmt_insert_query_ex.py

@@ -80,93 +80,9 @@ class TDTestCase:
         con=taos.connect(host=host, user=user, password=password, config=cfg ,port=port)
         print(con)
         return con
-
-    def test_stmt_insert_multi(self,conn):
-        # type: (TaosConnection) -> None
-
-        dbname = "pytest_taos_stmt_multi"
-        try:
-            conn.execute("drop database if exists %s" % dbname)
-            conn.execute("create database if not exists %s" % dbname)
-            conn.select_db(dbname)
-
-            conn.execute(
-                "create table if not exists log(ts timestamp, bo bool, nil tinyint, ti tinyint, si smallint, ii int,\
-                bi bigint, tu tinyint unsigned, su smallint unsigned, iu int unsigned, bu bigint unsigned, \
-                ff float, dd double, bb binary(100), nn nchar(100), tt timestamp)",
-            )
-            # conn.load_table_info("log")
-
-            start = datetime.now()
-            stmt = conn.statement("insert into log values(?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?)")
-
-            params = new_multi_binds(16)
-            params[0].timestamp((1626861392589, 1626861392590, 1626861392591))
-            params[1].bool((True, None, False))
-            params[2].tinyint([-128, -128, None]) # -128 is tinyint null
-            params[3].tinyint([0, 127, None])
-            params[4].smallint([3, None, 2])
-            params[5].int([3, 4, None])
-            params[6].bigint([3, 4, None])
-            params[7].tinyint_unsigned([3, 4, None])
-            params[8].smallint_unsigned([3, 4, None])
-            params[9].int_unsigned([3, 4, None])
-            params[10].bigint_unsigned([3, 4, None])
-            params[11].float([3, None, 1])
-            params[12].double([3, None, 1.2])
-            params[13].binary(["abc", "dddafadfadfadfadfa", None])
-            params[14].nchar(["涛思数据", None, "a long string with 中文字符"])
-            params[15].timestamp([None, None, 1626861392591])
-            # print(type(stmt))
-            stmt.bind_param_batch(params)
-            stmt.execute()
-            end = datetime.now()
-            print("elapsed time: ", end - start)
-            assert stmt.affected_rows == 3
-            
-            #query
-            querystmt=conn.statement("select ?,bu from log")
-            queryparam=new_bind_params(1)
-            print(type(queryparam))
-            queryparam[0].binary("ts")
-            querystmt.bind_param(queryparam)
-            querystmt.execute() 
-            result=querystmt.use_result()
-            # rows=result.fetch_all()
-            # print( querystmt.use_result())
-
-            # result = conn.query("select * from log")
-            rows=result.fetch_all()
-            # rows=result.fetch_all()
-            print(rows)
-            assert rows[1][0] == "ts"
-            assert rows[0][1] == 3
-
-            #query
-            querystmt1=conn.statement("select * from log where bu < ?")
-            queryparam1=new_bind_params(1)
-            print(type(queryparam1))
-            queryparam1[0].int(4)
-            querystmt1.bind_param(queryparam1)
-            querystmt1.execute() 
-            result1=querystmt1.use_result()
-            rows1=result1.fetch_all()
-            assert str(rows1[0][0]) == "2021-07-21 17:56:32.589000"
-            assert rows1[0][10] == 3
-
-
-            stmt.close()
-
-            # conn.execute("drop database if exists %s" % dbname)
-            conn.close()
-
-        except Exception as err:
-            # conn.execute("drop database if exists %s" % dbname)
-            conn.close()
-            raise err
         
     def test_stmt_set_tbname_tag(self,conn):
-        dbname = "pytest_taos_stmt_set_tbname_tag"
+        dbname = "stmt_set_tbname_tag"
         
         try:
             conn.execute("drop database if exists %s" % dbname)
@@ -174,16 +90,16 @@ class TDTestCase:
             conn.select_db(dbname)
             conn.execute("create table if not exists log(ts timestamp, bo bool, nil tinyint, ti tinyint, si smallint, ii int,\
                 bi bigint, tu tinyint unsigned, su smallint unsigned, iu int unsigned, bu bigint unsigned, \
-                ff float, dd double, bb binary(100), nn nchar(100), tt timestamp) tags (t1 timestamp, t2 bool,\
+                ff float, dd double, bb binary(100), nn nchar(100), tt timestamp , vc varchar(100)) tags (t1 timestamp, t2 bool,\
                 t3 tinyint, t4 tinyint, t5 smallint, t6 int, t7 bigint, t8 tinyint unsigned, t9 smallint unsigned, \
                 t10 int unsigned, t11 bigint unsigned, t12 float, t13 double, t14 binary(100), t15 nchar(100), t16 timestamp)")
             
             stmt = conn.statement("insert into ? using log tags (?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?) \
-                values (?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?)")
+                values (?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?)")
             tags = new_bind_params(16)
             tags[0].timestamp(1626861392589123, PrecisionEnum.Microseconds)
             tags[1].bool(True)
-            tags[2].null()
+            tags[2].bool(False)
             tags[3].tinyint(2)
             tags[4].smallint(3)
             tags[5].int(4)
@@ -198,7 +114,7 @@ class TDTestCase:
             tags[14].nchar("stmt")
             tags[15].timestamp(1626861392589, PrecisionEnum.Milliseconds)
             stmt.set_tbname_tags("tb1", tags)
-            params = new_multi_binds(16)
+            params = new_multi_binds(17)
             params[0].timestamp((1626861392589111,  1626861392590111, 1626861392591111))
             params[1].bool((True, None, False))
             params[2].tinyint([-128, -128, None]) # -128 is tinyint null
@@ -213,25 +129,32 @@ class TDTestCase:
             params[11].float([3, None, 1])
             params[12].double([3, None, 1.2])
             params[13].binary(["abc", "dddafadfadfadfadfa", None])
-            params[14].nchar(["涛思数据", None, "a? long string with 中文字符"])
+            params[14].nchar(["涛思数据", None, "a long string with 中文字符"])
             params[15].timestamp([None, None, 1626861392591])
-            
+            params[16].binary(["涛思数据16", None, "a long string with 中文-字符"])
+                     
             stmt.bind_param_batch(params)
             stmt.execute()
 
             assert stmt.affected_rows == 3
 
-            #query
+            #query all 
             querystmt1=conn.statement("select * from log where bu < ?")
             queryparam1=new_bind_params(1)
             print(type(queryparam1))
-            queryparam1[0].int(5)
+            queryparam1[0].int(10)
             querystmt1.bind_param(queryparam1)
             querystmt1.execute() 
             result1=querystmt1.use_result()
             rows1=result1.fetch_all()
-            print("1",rows1)
+            print(rows1[0])
+            print(rows1[1])
+            print(rows1[2])
+            assert str(rows1[0][0]) == "2021-07-21 17:56:32.589111"
+            assert rows1[0][10] == 3
+            assert rows1[1][10] == 4
 
+            #query: Numeric Functions
             querystmt2=conn.statement("select abs(?) from log where bu < ?")
             queryparam2=new_bind_params(2)
             print(type(queryparam2))
@@ -242,6 +165,11 @@ class TDTestCase:
             result2=querystmt2.use_result()
             rows2=result2.fetch_all()
             print("2",rows2)
+            assert rows2[0][0] == 5
+            assert rows2[1][0] == 5
+
+
+            #query: Numeric Functions and escapes
 
             querystmt3=conn.statement("select abs(?) from log where  nn= 'a? long string with 中文字符' ")
             queryparam3=new_bind_params(1)
@@ -252,9 +180,63 @@ class TDTestCase:
             result3=querystmt3.use_result()
             rows3=result3.fetch_all()
             print("3",rows3)
-            # assert str(rows1[0][0]) == "2021-07-21 17:56:32.589111"
-            # assert rows1[0][10] == 3
-            # assert rows1[1][10] == 4
+            assert rows3 == []
+
+            #query: string Functions
+
+            querystmt3=conn.statement("select CHAR_LENGTH(?) from log  ")
+            queryparam3=new_bind_params(1)
+            print(type(queryparam3))
+            queryparam3[0].binary('中文字符')
+            querystmt3.bind_param(queryparam3)
+            querystmt3.execute() 
+            result3=querystmt3.use_result()
+            rows3=result3.fetch_all()
+            print("4",rows3)
+            assert rows3[0][0] == 12, 'fourth case is failed'
+            assert rows3[1][0] == 12, 'fourth case is failed'
+
+            # #query: conversion Functions
+
+            # querystmt4=conn.statement("select cast( ? as bigint) from log  ")
+            # queryparam4=new_bind_params(1)
+            # print(type(queryparam4))
+            # queryparam4[0].binary('1232a')
+            # querystmt4.bind_param(queryparam4)
+            # querystmt4.execute() 
+            # result4=querystmt4.use_result()
+            # rows4=result4.fetch_all()
+            # print("5",rows4)
+            # assert rows4[0][0] == 1232
+            # assert rows4[1][0] == 1232
+
+            # querystmt4=conn.statement("select cast( ? as binary(10)) from log  ")
+            # queryparam4=new_bind_params(1)
+            # print(type(queryparam4))
+            # queryparam4[0].int(123)
+            # querystmt4.bind_param(queryparam4)
+            # querystmt4.execute() 
+            # result4=querystmt4.use_result()
+            # rows4=result4.fetch_all()
+            # print("6",rows4)
+            # assert rows4[0][0] == '123'
+            # assert rows4[1][0] == '123'
+
+            # #query: datatime Functions
+
+            # querystmt4=conn.statement(" select timediff('2021-07-21 17:56:32.590111',?,1s)  from log  ")
+            # queryparam4=new_bind_params(1)
+            # print(type(queryparam4))
+            # queryparam4[0].timestamp(1626861392591111)
+            # querystmt4.bind_param(queryparam4)
+            # querystmt4.execute() 
+            # result4=querystmt4.use_result()
+            # rows4=result4.fetch_all()
+            # print("7",rows4)
+            # assert rows4[0][0] == 1, 'seventh case is failed'
+            # assert rows4[1][0] == 1, 'seventh case is failed'
+            
+            
 
             # conn.execute("drop database if exists %s" % dbname)
             conn.close()
@@ -269,8 +251,6 @@ class TDTestCase:
         config = buildPath+ "../sim/dnode1/cfg/"
         host="localhost"
         connectstmt=self.newcon(host,config)
-        self.test_stmt_insert_multi(connectstmt)
-        connectstmt=self.newcon(host,config)
         self.test_stmt_set_tbname_tag(connectstmt)
 
         return 

tests/system-test/1-insert/test_stmt_muti_insert_query.py

@@ -124,7 +124,7 @@ class TDTestCase:
             print("elapsed time: ", end - start)
             assert stmt.affected_rows == 3
             
-            #query
+            #query 1
             querystmt=conn.statement("select ?,bu from log")
             queryparam=new_bind_params(1)
             print(type(queryparam))
@@ -141,8 +141,9 @@ class TDTestCase:
             print(rows)
             assert rows[1][0] == "ts"
             assert rows[0][1] == 3
+            assert rows[2][1] == None
 
-            #query
+            #query 2
             querystmt1=conn.statement("select * from log where bu < ?")
             queryparam1=new_bind_params(1)
             print(type(queryparam1))

tests/system-test/1-insert/test_stmt_set_tbname_tag.py

@@ -82,7 +82,7 @@ class TDTestCase:
         return con
         
     def test_stmt_set_tbname_tag(self,conn):
-        dbname = "pytest_taos_stmt_set_tbname_tag"
+        dbname = "stmt_set_tbname_tag"
         
         try:
             conn.execute("drop database if exists %s" % dbname)
@@ -99,7 +99,7 @@ class TDTestCase:
             tags = new_bind_params(16)
             tags[0].timestamp(1626861392589123, PrecisionEnum.Microseconds)
             tags[1].bool(True)
-            tags[2].null()
+            tags[2].bool(False)
             tags[3].tinyint(2)
             tags[4].smallint(3)
             tags[5].int(4)
@@ -114,7 +114,7 @@ class TDTestCase:
             tags[14].nchar("stmt")
             tags[15].timestamp(1626861392589, PrecisionEnum.Milliseconds)
             stmt.set_tbname_tags("tb1", tags)
-            params = new_multi_binds(16)
+            params = new_multi_binds(17)
             params[0].timestamp((1626861392589111,  1626861392590111, 1626861392591111))
             params[1].bool((True, None, False))
             params[2].tinyint([-128, -128, None]) # -128 is tinyint null
@@ -129,28 +129,153 @@ class TDTestCase:
             params[11].float([3, None, 1])
             params[12].double([3, None, 1.2])
             params[13].binary(["abc", "dddafadfadfadfadfa", None])
-            params[14].nchar(["涛思数据", None, "a long string with 中文字符"])
+            params[14].nchar(["涛思数据", None, "a long string with 中文?字符"])
             params[15].timestamp([None, None, 1626861392591])
-            params[16].binary(["涛思数据16", None, "a long string with 中文-字符"])
+            params[16].binary(["涛思数据16", None, None])
                      
             stmt.bind_param_batch(params)
             stmt.execute()
 
             assert stmt.affected_rows == 3
 
-            #query
+            #query all 
             querystmt1=conn.statement("select * from log where bu < ?")
             queryparam1=new_bind_params(1)
             print(type(queryparam1))
-            queryparam1[0].int(5)
+            queryparam1[0].int(10)
             querystmt1.bind_param(queryparam1)
             querystmt1.execute() 
             result1=querystmt1.use_result()
             rows1=result1.fetch_all()
-            print(rows1)
-            # assert str(rows1[0][0]) == "2021-07-21 17:56:32.589111"
-            # assert rows1[0][10] == 3
-            # assert rows1[1][10] == 4
+            print(rows1[0])
+            print(rows1[1])
+            print(rows1[2])
+            assert str(rows1[0][0]) == "2021-07-21 17:56:32.589111"
+            assert rows1[0][10] == 3
+            assert rows1[1][10] == 4
+
+            #query: Numeric Functions
+            querystmt2=conn.statement("select abs(?) from log where bu < ?")
+            queryparam2=new_bind_params(2)
+            print(type(queryparam2))
+            queryparam2[0].int(5)
+            queryparam2[1].int(5)
+            querystmt2.bind_param(queryparam2)
+            querystmt2.execute() 
+            result2=querystmt2.use_result()
+            rows2=result2.fetch_all()
+            print("2",rows2)
+            assert rows2[0][0] == 5
+            assert rows2[1][0] == 5
+
+
+            #query: Numeric Functions and escapes
+
+            querystmt3=conn.statement("select abs(?) from log where  nn= 'a? long string with 中文字符' ")
+            queryparam3=new_bind_params(1)
+            print(type(queryparam3))
+            queryparam3[0].int(5)
+            querystmt3.bind_param(queryparam3)
+            querystmt3.execute() 
+            result3=querystmt3.use_result()
+            rows3=result3.fetch_all()
+            print("3",rows3)
+            assert rows3 == []
+
+            # #query: string Functions
+
+            # querystmt3=conn.statement("select CHAR_LENGTH(?) from log  ")
+            # queryparam3=new_bind_params(1)
+            # print(type(queryparam3))
+            # queryparam3[0].binary('中文字符')
+            # querystmt3.bind_param(queryparam3)
+            # querystmt3.execute() 
+            # result3=querystmt3.use_result()
+            # rows3=result3.fetch_all()
+            # print("4",rows3)
+            # assert rows3[0][0] == 12, 'fourth case is failed'
+            # assert rows3[1][0] == 12, 'fourth case is failed'
+
+            # #query: conversion Functions
+
+            # querystmt4=conn.statement("select cast( ? as bigint) from log  ")
+            # queryparam4=new_bind_params(1)
+            # print(type(queryparam4))
+            # queryparam4[0].binary('1232a')
+            # querystmt4.bind_param(queryparam4)
+            # querystmt4.execute() 
+            # result4=querystmt4.use_result()
+            # rows4=result4.fetch_all()
+            # print("5",rows4)
+            # assert rows4[0][0] == 1232
+            # assert rows4[1][0] == 1232
+
+            # querystmt4=conn.statement("select cast( ? as binary(10)) from log  ")
+            # queryparam4=new_bind_params(1)
+            # print(type(queryparam4))
+            # queryparam4[0].int(123)
+            # querystmt4.bind_param(queryparam4)
+            # querystmt4.execute() 
+            # result4=querystmt4.use_result()
+            # rows4=result4.fetch_all()
+            # print("6",rows4)
+            # assert rows4[0][0] == '123'
+            # assert rows4[1][0] == '123'
+
+            # #query: datatime Functions
+
+            # querystmt4=conn.statement(" select timediff('2021-07-21 17:56:32.590111',?,1s)  from log  ")
+            # queryparam4=new_bind_params(1)
+            # print(type(queryparam4))
+            # queryparam4[0].timestamp(1626861392591111)
+            # querystmt4.bind_param(queryparam4)
+            # querystmt4.execute() 
+            # result4=querystmt4.use_result()
+            # rows4=result4.fetch_all()
+            # print("7",rows4)
+            # assert rows4[0][0] == 1, 'seventh case is failed'
+            # assert rows4[1][0] == 1, 'seventh case is failed'
+            
+            #query: aggregate Functions
+
+            querystmt4=conn.statement(" select count(?)  from log  ")
+            queryparam4=new_bind_params(1)
+            print(type(queryparam4))
+            queryparam4[0].int(123)
+            querystmt4.bind_param(queryparam4)
+            querystmt4.execute() 
+            result4=querystmt4.use_result()
+            rows4=result4.fetch_all()
+            print("8",rows4)
+            assert rows4[0][0] == 3, ' 8 case is failed'
+
+            #query: selector Functions 9
+
+            querystmt4=conn.statement("  select bottom(bu,?)  from log group by bu ;  ")
+            queryparam4=new_bind_params(1)
+            print(type(queryparam4))
+            queryparam4[0].int(2)
+            querystmt4.bind_param(queryparam4)
+            querystmt4.execute() 
+            result4=querystmt4.use_result()
+            rows4=result4.fetch_all()
+            print("9",rows4)
+            assert rows4[0][0] == 4, ' 9 case is failed'
+            assert rows4[1][0] == 3, ' 9 case is failed'
+
+            # #query: time-series specific Functions 10
+
+            querystmt4=conn.statement("  select twa(?)  from log;  ")
+            queryparam4=new_bind_params(1)
+            print(type(queryparam4))
+            queryparam4[0].int(15)
+            querystmt4.bind_param(queryparam4)
+            querystmt4.execute() 
+            result4=querystmt4.use_result()
+            rows4=result4.fetch_all()
+            print("10",rows4)
+            assert rows4[0][0] == 15, ' 10 case is failed'
+
 
             # conn.execute("drop database if exists %s" % dbname)
             conn.close()