diff --git a/src/client/src/tscParseInsert.c b/src/client/src/tscParseInsert.c
index 125e28f17778baead6b2be8d8409f2259d7d77e5..0750abdf6ff991d8c19eb63b77c5ba9ff9ddfb03 100644
--- a/src/client/src/tscParseInsert.c
+++ b/src/client/src/tscParseInsert.c
@@ -762,35 +762,32 @@ void tscSortRemoveDataBlockDupRowsRaw(STableDataBlocks *dataBuf) {
   if (!dataBuf->ordered) {
     char *pBlockData = pBlocks->data;
     qsort(pBlockData, pBlocks->numOfRows, dataBuf->rowSize, rowDataCompar);
+    dataBuf->ordered = true;
 
-    int32_t i = 0;
-    int32_t j = 1;
-
-    while (j < pBlocks->numOfRows) {
-      TSKEY ti = *(TSKEY *)(pBlockData + dataBuf->rowSize * i);
-      TSKEY tj = *(TSKEY *)(pBlockData + dataBuf->rowSize * j);
+    if(tsClientMerge) {
+      int32_t i = 0;
+      int32_t j = 1;
+      while (j < pBlocks->numOfRows) {
+        TSKEY ti = *(TSKEY *)(pBlockData + dataBuf->rowSize * i);
+        TSKEY tj = *(TSKEY *)(pBlockData + dataBuf->rowSize * j);
 
-      if (ti == tj) {
-        if (dataBuf->pTableMeta && dataBuf->pTableMeta->tableInfo.update != TD_ROW_DISCARD_UPDATE) {
-          memmove(pBlockData + dataBuf->rowSize * i, pBlockData + dataBuf->rowSize * j, dataBuf->rowSize);
+        if (ti == tj) {
+          if (dataBuf->pTableMeta && dataBuf->pTableMeta->tableInfo.update != TD_ROW_DISCARD_UPDATE) {
+            memmove(pBlockData + dataBuf->rowSize * i, pBlockData + dataBuf->rowSize * j, dataBuf->rowSize);
+          }
+          ++j;
+          continue;
         }
 
+        int32_t nextPos = (++i);
+        if (nextPos != j) {
+          memmove(pBlockData + dataBuf->rowSize * nextPos, pBlockData + dataBuf->rowSize * j, dataBuf->rowSize);
+        }
         ++j;
-        continue;
-      }
-
-      int32_t nextPos = (++i);
-      if (nextPos != j) {
-        memmove(pBlockData + dataBuf->rowSize * nextPos, pBlockData + dataBuf->rowSize * j, dataBuf->rowSize);
-      }
-
-      ++j;
+      }      
+      pBlocks->numOfRows = i + 1;
+      dataBuf->size = sizeof(SSubmitBlk) + dataBuf->rowSize * pBlocks->numOfRows;
     }
-
-    dataBuf->ordered = true;
-
-    pBlocks->numOfRows = i + 1;
-    dataBuf->size = sizeof(SSubmitBlk) + dataBuf->rowSize * pBlocks->numOfRows;
   }
 
   dataBuf->prevTS = INT64_MIN;
@@ -836,32 +833,33 @@ int tscSortRemoveDataBlockDupRows(STableDataBlocks *dataBuf, SBlockKeyInfo *pBlk
   if (!dataBuf->ordered) {
     pBlkKeyTuple = pBlkKeyInfo->pKeyTuple;
     qsort(pBlkKeyTuple, nRows, sizeof(SBlockKeyTuple), rowDataCompar);
+    dataBuf->ordered = true;
 
-    pBlkKeyTuple = pBlkKeyInfo->pKeyTuple;
-    int32_t i = 0;
-    int32_t j = 1;
-    while (j < nRows) {
-      TSKEY ti = (pBlkKeyTuple + i)->skey;
-      TSKEY tj = (pBlkKeyTuple + j)->skey;
-
-      if (ti == tj) {
-        if (dataBuf->pTableMeta && dataBuf->pTableMeta->tableInfo.update != TD_ROW_DISCARD_UPDATE) {
-          memmove(pBlkKeyTuple + i, pBlkKeyTuple + j, sizeof(SBlockKeyTuple));
+    if(tsClientMerge) {
+      pBlkKeyTuple = pBlkKeyInfo->pKeyTuple;
+      int32_t i = 0;
+      int32_t j = 1;
+      while (j < nRows) {
+        TSKEY ti = (pBlkKeyTuple + i)->skey;
+        TSKEY tj = (pBlkKeyTuple + j)->skey;
+
+        if (ti == tj) {
+          if (dataBuf->pTableMeta && dataBuf->pTableMeta->tableInfo.update != TD_ROW_DISCARD_UPDATE) {
+            memmove(pBlkKeyTuple + i, pBlkKeyTuple + j, sizeof(SBlockKeyTuple));
+          }
+
+          ++j;
+          continue;
         }
 
+        int32_t nextPos = (++i);
+        if (nextPos != j) {
+          memmove(pBlkKeyTuple + nextPos, pBlkKeyTuple + j, sizeof(SBlockKeyTuple));
+        }
         ++j;
-        continue;
-      }
-
-      int32_t nextPos = (++i);
-      if (nextPos != j) {
-        memmove(pBlkKeyTuple + nextPos, pBlkKeyTuple + j, sizeof(SBlockKeyTuple));
       }
-      ++j;
-    }
-
-    dataBuf->ordered = true;
-    pBlocks->numOfRows = i + 1;
+      pBlocks->numOfRows = i + 1;
+    } 
   }
 
   dataBuf->size = sizeof(SSubmitBlk) + pBlocks->numOfRows * extendedRowSize;
diff --git a/src/client/src/tscSQLParser.c b/src/client/src/tscSQLParser.c
index 2103c5c8ededaee59b22f9a024b71d57ff241703..4cfec94f704d2821e9400861a181369743067185 100644
--- a/src/client/src/tscSQLParser.c
+++ b/src/client/src/tscSQLParser.c
@@ -5423,10 +5423,14 @@ int32_t validateFillNode(SSqlCmd* pCmd, SQueryInfo* pQueryInfo, SSqlNode* pSqlNo
   const char* msg4 = "illegal value or data overflow";
   const char* msg5 = "fill only available for interval query";
   const char* msg6 = "not supported function now";
+  const char* msg7 = "join query not supported fill operation";
 
   if ((!isTimeWindowQuery(pQueryInfo)) && (!tscIsPointInterpQuery(pQueryInfo))) {
     return invalidOperationMsg(tscGetErrorMsgPayload(pCmd), msg5);
   }
+  if(QUERY_IS_JOIN_QUERY(pQueryInfo->type)) {
+    return invalidOperationMsg(tscGetErrorMsgPayload(pCmd), msg7);
+  }
 
   /*
    * fill options are set at the end position, when all columns are set properly
diff --git a/src/common/inc/tglobal.h b/src/common/inc/tglobal.h
index a384cf6e70485b8f7d1b06b7f7e86ba92776b547..8a6cc9ddb5a4425e3d7bd10bb7dd5dda0712b0c3 100644
--- a/src/common/inc/tglobal.h
+++ b/src/common/inc/tglobal.h
@@ -209,6 +209,8 @@ extern int32_t wDebugFlag;
 extern int32_t cqDebugFlag;
 extern int32_t debugFlag;
 
+extern int8_t tsClientMerge;
+
 #ifdef TD_TSZ
 // lossy 
 extern char lossyColumns[];
@@ -219,6 +221,7 @@ extern uint32_t curRange;
 extern char Compressor[];
 #endif
 
+
 typedef struct {
   char dir[TSDB_FILENAME_LEN];
   int  level;
diff --git a/src/common/src/tglobal.c b/src/common/src/tglobal.c
index 876eb4fa92bada1d22e0e43b6d6532d31a0b913c..f1a7e4f7fd12a85ed1b8b503db35fba7211fc788 100644
--- a/src/common/src/tglobal.c
+++ b/src/common/src/tglobal.c
@@ -254,6 +254,8 @@ int32_t tsdbDebugFlag = 131;
 int32_t cqDebugFlag = 131;
 int32_t fsDebugFlag = 135;
 
+int8_t tsClientMerge = 0;
+
 #ifdef TD_TSZ
 //
 // lossy compress 6
@@ -1581,6 +1583,16 @@ static void doInitGlobalConfig(void) {
   cfg.unitType = TAOS_CFG_UTYPE_NONE;
   taosInitConfigOption(cfg);
 
+  cfg.option = "clientMerge";
+  cfg.ptr = &tsClientMerge;
+  cfg.valType = TAOS_CFG_VTYPE_INT8;
+  cfg.cfgType = TSDB_CFG_CTYPE_B_CONFIG | TSDB_CFG_CTYPE_B_SHOW;
+  cfg.minValue = 0;
+  cfg.maxValue = 1;
+  cfg.ptrLength = 1;
+  cfg.unitType = TAOS_CFG_UTYPE_NONE;
+  taosInitConfigOption(cfg);
+
 #ifdef TD_TSZ
   // lossy compress
   cfg.option = "lossyColumns";
diff --git a/src/query/src/qExecutor.c b/src/query/src/qExecutor.c
index ae4db6bdb8a7887e7adc7bf15354a104f1a903fa..1d90fcf1ac2d021e5d6063ea5d5b1aa03c8525e2 100644
--- a/src/query/src/qExecutor.c
+++ b/src/query/src/qExecutor.c
@@ -1296,9 +1296,6 @@ static void projectApplyFunctions(SQueryRuntimeEnv *pRuntimeEnv, SQLFunctionCtx
     if (pCtx[k].currentStage == MERGE_STAGE) {
       pCtx[k].order = TSDB_ORDER_ASC;
     }
-
-    pCtx[k].startTs = pQueryAttr->window.skey;
-
     if (pCtx[k].functionId < 0) {
       // load the script and exec
       SUdfInfo* pUdfInfo = pRuntimeEnv->pUdfInfo;
diff --git a/src/util/inc/tconfig.h b/src/util/inc/tconfig.h
index 5bd69aece2ffa2166c268b437549c95a30608383..d6a1cb113a43874ee77b096e75959e02e4aa6747 100644
--- a/src/util/inc/tconfig.h
+++ b/src/util/inc/tconfig.h
@@ -20,7 +20,7 @@
 extern "C" {
 #endif
 
-#define TSDB_CFG_MAX_NUM    119  
+#define TSDB_CFG_MAX_NUM    120  
 #define TSDB_CFG_PRINT_LEN  23
 #define TSDB_CFG_OPTION_LEN 24
 #define TSDB_CFG_VALUE_LEN  41
diff --git a/src/util/src/tskiplist.c b/src/util/src/tskiplist.c
index 98fd9c094cba3e779c9f203fdacc548a3bda5ef4..2bdaeb07a16a638d75c1c5a8b20763e5c660a05f 100644
--- a/src/util/src/tskiplist.c
+++ b/src/util/src/tskiplist.c
@@ -144,7 +144,6 @@ void tSkipListPutBatchByIter(SSkipList *pSkipList, void *iter, iter_next_fn_t it
 
   // backward to put the first data
   hasDup = tSkipListGetPosToPut(pSkipList, backward, pData);
-
   tSkipListPutImpl(pSkipList, pData, backward, false, hasDup);
 
   for (int level = 0; level < pSkipList->maxLevel; level++) {
@@ -163,7 +162,12 @@ void tSkipListPutBatchByIter(SSkipList *pSkipList, void *iter, iter_next_fn_t it
       for (int i = 0; i < pSkipList->maxLevel; i++) {
         forward[i] = SL_NODE_GET_BACKWARD_POINTER(pSkipList->pTail, i);
       }
+    } else if(compare == 0) {
+      // same need special deal
+      forward[0] = SL_NODE_GET_BACKWARD_POINTER(SL_NODE_GET_BACKWARD_POINTER(pSkipList->pTail,0),0);
+      hasDup = true;
     } else {
+      SSkipListNode *p  = NULL;
       SSkipListNode *px = pSkipList->pHead;
       for (int i = pSkipList->maxLevel - 1; i >= 0; --i) {
         if (i < pSkipList->level) {
@@ -175,19 +179,29 @@ void tSkipListPutBatchByIter(SSkipList *pSkipList, void *iter, iter_next_fn_t it
             }
           }
 
-          SSkipListNode *p = SL_NODE_GET_FORWARD_POINTER(px, i);
+          // if px not head , must compare with px
+          if(px == pSkipList->pHead) {
+            p = SL_NODE_GET_FORWARD_POINTER(px, i);
+          } else {
+            p = px;
+          }
           while (p != pSkipList->pTail) {
             pKey = SL_GET_NODE_KEY(pSkipList, p);
 
             compare = pSkipList->comparFn(pKey, pDataKey);
             if (compare >= 0) {
-              if (compare == 0 && !hasDup) hasDup = true;
+              if (compare == 0) {
+                hasDup = true;
+                forward[0] = SL_NODE_GET_BACKWARD_POINTER(p, 0);
+              }
               break;
             } else {
               px = p;
               p = SL_NODE_GET_FORWARD_POINTER(px, i);
             }
           }
+          // if found duplicate, immediately break, needn't continue to loop set rest forward[i] value
+          if(hasDup) break;
         }
 
         forward[i] = px;
diff --git a/tests/pytest/query/nestedQuery/nestedQuery.py b/tests/pytest/query/nestedQuery/nestedQuery.py
index 453ee8f53975509c318486242c634d3b60de4992..545f6429e825c468bdb07524329d6ea49944e379 100755
--- a/tests/pytest/query/nestedQuery/nestedQuery.py
+++ b/tests/pytest/query/nestedQuery/nestedQuery.py
@@ -1714,7 +1714,6 @@ class TDTestCase:
             sql += " from regular_table_1 t1, regular_table_2 t2 where t1.ts = t2.ts and   "
             sql += "%s " % random.choice(q_u_where)
             sql += "%s " % random.choice(session_u_where)
-            sql += "%s " % random.choice(fill_where)
             sql += "%s " % random.choice(order_u_where)
             sql += "%s " % random.choice(limit_u_where)
             sql += ") "
@@ -1731,7 +1730,6 @@ class TDTestCase:
             sql += " from regular_table_1 t1, regular_table_2 t2 where t1.ts = t2.ts and   "
             sql += "%s " % random.choice(q_u_or_where)
             sql += "%s " % random.choice(session_u_where)
-            sql += "%s " % random.choice(fill_where)
             sql += "%s " % random.choice(order_u_where)
             sql += "%s " % random.choice(limit_u_where)
             sql += ") "
@@ -1767,7 +1765,6 @@ class TDTestCase:
             sql += " from table_1 t1, regular_table_2 t2 where t1.ts = t2.ts and  "
             sql += "%s " % random.choice(q_u_where)
             sql += "%s " % random.choice(session_u_where)
-            sql += "%s " % random.choice(fill_where)
             sql += "%s " % random.choice(order_u_where)
             sql += "%s " % random.choice(limit_u_where)
             sql += ") "
@@ -1784,7 +1781,6 @@ class TDTestCase:
             sql += " from table_1 t1, regular_table_2 t2 where t1.ts = t2.ts and  "
             sql += "%s " % random.choice(q_u_or_where)
             sql += "%s " % random.choice(session_u_where)
-            sql += "%s " % random.choice(fill_where)
             sql += "%s " % random.choice(order_u_where)
             sql += "%s " % random.choice(limit_u_where)
             sql += ") "
@@ -1818,7 +1814,6 @@ class TDTestCase:
             sql += " from stable_1 t1, stable_2 t2 where t1.ts = t2.ts and   "
             sql += "%s " % random.choice(t_join_where)
             sql += "%s " % random.choice(session_u_where)
-            sql += "%s " % random.choice(fill_where)
             sql += "%s " % random.choice(order_u_where)
             sql += "%s " % random.choice(limit_u_where)
             sql += ") "
@@ -1835,7 +1830,6 @@ class TDTestCase:
             sql += " from stable_1 t1, stable_2 t2 where t1.ts = t2.ts and   "
             sql += "%s " % random.choice(qt_u_or_where)
             sql += "%s " % random.choice(session_u_where)
-            sql += "%s " % random.choice(fill_where)
             sql += "%s " % random.choice(order_u_where)
             sql += "%s " % random.choice(limit_u_where)
             sql += ") "
@@ -2015,7 +2009,6 @@ class TDTestCase:
             sql += " from stable_1 t1 , stable_2 t2 where t1.ts = t2.ts and  "
             sql += "%s and " % random.choice(t_join_where)
             sql += "%s " % random.choice(interp_where_j)  
-            sql += "%s " % random.choice(fill_where)
             sql += "%s " % random.choice(order_u_where)
             sql += "%s " % random.choice(limit_u_where)
             sql += ") "
@@ -2032,7 +2025,6 @@ class TDTestCase:
             sql += " from stable_1 t1 , stable_2 t2 where t1.ts = t2.ts and  "
             sql += "%s and " % random.choice(qt_u_or_where)
             sql += "%s " % random.choice(interp_where_j)  
-            sql += "%s " % random.choice(fill_where)
             sql += "%s " % random.choice(order_u_where)
             sql += "%s " % random.choice(limit_u_where)
             sql += ") "
@@ -2065,7 +2057,6 @@ class TDTestCase:
             sql += " from table_0 t1, table_1 t2 where t1.ts = t2.ts and    "             
             #sql += "%s and " % random.choice(t_join_where)
             sql += "%s " % interp_where_j[random.randint(0,5)]
-            sql += "%s " % random.choice(fill_where)
             sql += "%s " % random.choice(order_u_where)
             sql += "%s " % random.choice(limit_u_where)
             sql += ") "
@@ -2116,7 +2107,6 @@ class TDTestCase:
             sql += " from regular_table_1 t1, regular_table_2 t2 where t1.ts = t2.ts and   "
             #sql += "%s " % random.choice(interp_where_j) 
             sql += "%s " % interp_where_j[random.randint(0,5)] 
-            sql += "%s " % random.choice(fill_where)
             sql += "%s " % random.choice(order_u_where)
             sql += "%s " % random.choice(limit_u_where)
             sql += ") "
diff --git a/tests/script/general/parser/join.sim b/tests/script/general/parser/join.sim
index e2132589bd3a54d42e683094f184b3a4a4932f71..36deea0371486394125355cc92a9785764610569 100644
--- a/tests/script/general/parser/join.sim
+++ b/tests/script/general/parser/join.sim
@@ -445,7 +445,7 @@ if $rows != $val then
 endi
 
 print ================>TD-5600
-sql select first(join_tb0.c8),first(join_tb0.c9) from join_tb1 , join_tb0 where join_tb1.ts = join_tb0.ts and join_tb1.ts <= 100002 and join_tb1.ts>=100000 interval(1s) fill(linear);
+sql select first(join_tb0.c8),first(join_tb0.c9) from join_tb1 , join_tb0 where join_tb1.ts = join_tb0.ts and join_tb1.ts <= 100002 and join_tb1.ts>=100000 interval(1s);
 
 
 #===============================================================