Merge pull request #21405 from taosdata/szhou/scl-flt

enhance: scalar filter improvement

include/common/tglobal.h

@@ -180,6 +180,7 @@ extern int32_t tsRpcRetryInterval;
 extern bool    tsDisableStream;
 extern int64_t tsStreamBufferSize;
 extern int64_t tsCheckpointInterval;
+extern bool    tsFilterScalarMode;
 
 // #define NEEDTO_COMPRESSS_MSG(size) (tsCompressMsgSize != -1 && (size) > tsCompressMsgSize)
 

source/common/src/tglobal.c

@@ -208,6 +208,7 @@ char    tsUdfdLdLibPath[512] = "";
 bool    tsDisableStream = false;
 int64_t tsStreamBufferSize = 128 * 1024 * 1024;
 int64_t tsCheckpointInterval = 3 * 60 * 60 * 1000;
+bool    tsFilterScalarMode = false;
 
 #ifndef _STORAGE
 int32_t taosSetTfsCfg(SConfig *pCfg) {
@@ -522,6 +523,8 @@ static int32_t taosAddServerCfg(SConfig *pCfg) {
 
   if (cfgAddInt32(pCfg, "cacheLazyLoadThreshold", tsCacheLazyLoadThreshold, 0, 100000, 0) != 0) return -1;
 
+  if (cfgAddBool(pCfg, "filterScalarMode", tsFilterScalarMode, 0) != 0) return -1;
+
   GRANT_CFG_ADD;
   return 0;
 }
@@ -898,6 +901,8 @@ static int32_t taosSetServerCfg(SConfig *pCfg) {
   tsStreamBufferSize = cfgGetItem(pCfg, "streamBufferSize")->i64;
   tsCheckpointInterval = cfgGetItem(pCfg, "checkpointInterval")->i64;
 
+  tsFilterScalarMode = cfgGetItem(pCfg, "filterScalarMode")->bval;
+  
   GRANT_CFG_GET;
   return 0;
 }

source/libs/executor/src/executorInt.c

@@ -573,8 +573,18 @@ void extractQualifiedTupleByFilterResult(SSDataBlock* pBlock, const SColumnInfoD
           if (colDataIsNull_var(pDst, j)) {
             colDataSetNull_var(pDst, numOfRows);
           } else {
+            // fix address sanitizer error. p1 may point to memory that will change during realloc of colDataSetVal, first copy it to p2
             char* p1 = colDataGetVarData(pDst, j);
-            colDataSetVal(pDst, numOfRows, p1, false);
+            int32_t len = 0;
+            if (pDst->info.type == TSDB_DATA_TYPE_JSON) {
+              len = getJsonValueLen(p1);
+            } else {
+              len = varDataTLen(p1);
+            }
+            char* p2 = taosMemoryMalloc(len);
+            memcpy(p2, p1, len);
+            colDataSetVal(pDst, numOfRows, p2, false);
+            taosMemoryFree(p2);
           }
           numOfRows += 1;
           j += 1;

source/libs/function/src/udfd.c

@@ -600,9 +600,9 @@ SUdf *udfdGetOrCreateUdf(const char *udfName) {
       return udf;
     } else {
       (*pUdfHash)->expired = true;
-      taosHashRemove(global.udfsHash, udfName, strlen(udfName));
       fnInfo("udfd expired, check for new version. existing udf %s udf version %d, udf created time %" PRIx64,
              (*pUdfHash)->name, (*pUdfHash)->version, (*pUdfHash)->createdTime);
+      taosHashRemove(global.udfsHash, udfName, strlen(udfName));
     }
   }
 

source/libs/scalar/inc/filterInt.h

@@ -227,8 +227,10 @@ typedef struct SFltTreeStat {
   SFilterInfo *info;
 } SFltTreeStat;
 
+
 typedef struct SFltScalarCtx {
   SNode *node;
+  SArray* fltSclRange;
 } SFltScalarCtx;
 
 typedef struct SFltBuildGroupCtx {
@@ -237,6 +239,11 @@ typedef struct SFltBuildGroupCtx {
   int32_t      code;
 } SFltBuildGroupCtx;
 
+typedef struct {
+  SColumnNode *colNode;
+  SArray      *points;
+} SFltSclColumnRange;
+
 struct SFilterInfo {
   bool              scalarMode;
   SFltScalarCtx     sclCtx;

source/libs/scalar/src/sclvector.c

@@ -1791,7 +1791,11 @@ void vectorNotMatch(SScalarParam *pLeft, SScalarParam *pRight, SScalarParam *pOu
 void vectorIsNull(SScalarParam *pLeft, SScalarParam *pRight, SScalarParam *pOut, int32_t _ord) {
   for (int32_t i = 0; i < pLeft->numOfRows; ++i) {
     int8_t v = IS_HELPER_NULL(pLeft->columnData, i) ? 1 : 0;
+    if (v) {
+      ++pOut->numOfQualified;
+    }
     colDataSetInt8(pOut->columnData, i, &v);
+    colDataClearNull_f(pOut->columnData->nullbitmap, i);
   }
   pOut->numOfRows = pLeft->numOfRows;
 }
@@ -1799,7 +1803,11 @@ void vectorIsNull(SScalarParam *pLeft, SScalarParam *pRight, SScalarParam *pOut,
 void vectorNotNull(SScalarParam *pLeft, SScalarParam *pRight, SScalarParam *pOut, int32_t _ord) {
   for (int32_t i = 0; i < pLeft->numOfRows; ++i) {
     int8_t v = IS_HELPER_NULL(pLeft->columnData, i) ? 0 : 1;
+    if (v) {
+      ++pOut->numOfQualified;
+    }
     colDataSetInt8(pOut->columnData, i, &v);
+    colDataClearNull_f(pOut->columnData->nullbitmap, i);
   }
   pOut->numOfRows = pLeft->numOfRows;
 }
@@ -1812,6 +1820,13 @@ void vectorIsTrue(SScalarParam *pLeft, SScalarParam *pRight, SScalarParam *pOut,
       colDataSetInt8(pOut->columnData, i, &v);
       colDataClearNull_f(pOut->columnData->nullbitmap, i);
     }
+    {
+      bool v = false;
+      GET_TYPED_DATA(v, bool, pOut->columnData->info.type, colDataGetData(pOut->columnData, i));
+      if (v) {
+        ++pOut->numOfQualified;
+      }
+    }
   }
   pOut->columnData->hasNull = false;
 }
@@ -1851,7 +1866,9 @@ void vectorJsonContains(SScalarParam *pLeft, SScalarParam *pRight, SScalarParam
       char *pLeftData = colDataGetVarData(pLeft->columnData, i);
       getJsonValue(pLeftData, jsonKey, &isExist);
     }
-
+    if (isExist) {
+      ++pOut->numOfQualified;
+    }
     colDataSetVal(pOutputCol, i, (const char *)(&isExist), false);
   }
   taosMemoryFree(jsonKey);

tests/parallel_test/cases.task

@@ -1274,6 +1274,7 @@
 ,,y,script,./test.sh -f tsim/parser/columnValue_tinyint.sim
 ,,y,script,./test.sh -f tsim/parser/columnValue_unsign.sim
 ,,y,script,./test.sh -f tsim/parser/condition.sim
+,,y,script,./test.sh -f tsim/parser/condition_scl.sim
 ,,y,script,./test.sh -f tsim/parser/constCol.sim
 ,,y,script,./test.sh -f tsim/parser/create_db.sim
 ,,y,script,./test.sh -f tsim/parser/create_mt.sim
@@ -1353,6 +1354,7 @@
 ,,y,script,./test.sh -f tsim/query/event.sim
 ,,y,script,./test.sh -f tsim/query/forceFill.sim
 ,,y,script,./test.sh -f tsim/query/emptyTsRange.sim
+,,y,script,./test.sh -f tsim/query/emptyTsRange_scl.sim
 ,,y,script,./test.sh -f tsim/query/partitionby.sim
 ,,y,script,./test.sh -f tsim/query/tableCount.sim 
 ,,y,script,./test.sh -f tsim/query/tag_scan.sim 

tests/script/tsim/parser/condition_scl.sim

+system sh/stop_dnodes.sh
+system sh/deploy.sh -n dnode1 -i 1
+system sh/cfg.sh -n dnode1 -c filterScalarMode -v 1
+system sh/exec.sh -n dnode1 -s start
+sql connect
+
+sql drop database if exists cdb
+sql create database if not exists cdb
+sql use cdb
+sql create table stb1 (ts timestamp, c1 int, c2 float, c3 bigint, c4 smallint, c5 tinyint, c6 double, c7 bool, c8 binary(10), c9 nchar(9)) TAGS(t1 int, t2 binary(10), t3 double)
+sql create table tb1 using stb1 tags(1,'1',1.0)
+sql create table tb2 using stb1 tags(2,'2',2.0)
+sql create table tb3 using stb1 tags(3,'3',3.0)
+sql create table tb4 using stb1 tags(4,'4',4.0)
+sql create table tb5 using stb1 tags(5,'5',5.0)
+sql create table tb6 using stb1 tags(6,'6',6.0)
+
+sql insert into tb1 values ('2021-05-05 18:19:00',1,1.0,1,1,1,1.0,true ,'1','1')
+sql insert into tb1 values ('2021-05-05 18:19:01',2,2.0,2,2,2,2.0,true ,'2','2')
+sql insert into tb1 values ('2021-05-05 18:19:02',3,3.0,3,3,3,3.0,false,'3','3')
+sql insert into tb1 values ('2021-05-05 18:19:03',4,4.0,4,4,4,4.0,false,'4','4')
+sql insert into tb1 values ('2021-05-05 18:19:04',11,11.0,11,11,11,11.0,true ,'11','11')
+sql insert into tb1 values ('2021-05-05 18:19:05',12,12.0,12,12,12,12.0,true ,'12','12')
+sql insert into tb1 values ('2021-05-05 18:19:06',13,13.0,13,13,13,13.0,false,'13','13')
+sql insert into tb1 values ('2021-05-05 18:19:07',14,14.0,14,14,14,14.0,false,'14','14')
+sql insert into tb2 values ('2021-05-05 18:19:08',21,21.0,21,21,21,21.0,true ,'21','21')
+sql insert into tb2 values ('2021-05-05 18:19:09',22,22.0,22,22,22,22.0,true ,'22','22')
+sql insert into tb2 values ('2021-05-05 18:19:10',23,23.0,23,23,23,23.0,false,'23','23')
+sql insert into tb2 values ('2021-05-05 18:19:11',24,24.0,24,24,24,24.0,false,'24','24')
+sql insert into tb3 values ('2021-05-05 18:19:12',31,31.0,31,31,31,31.0,true ,'31','31')
+sql insert into tb3 values ('2021-05-05 18:19:13',32,32.0,32,32,32,32.0,true ,'32','32')
+sql insert into tb3 values ('2021-05-05 18:19:14',33,33.0,33,33,33,33.0,false,'33','33')
+sql insert into tb3 values ('2021-05-05 18:19:15',34,34.0,34,34,34,34.0,false,'34','34')
+sql insert into tb4 values ('2021-05-05 18:19:16',41,41.0,41,41,41,41.0,true ,'41','41')
+sql insert into tb4 values ('2021-05-05 18:19:17',42,42.0,42,42,42,42.0,true ,'42','42')
+sql insert into tb4 values ('2021-05-05 18:19:18',43,43.0,43,43,43,43.0,false,'43','43')
+sql insert into tb4 values ('2021-05-05 18:19:19',44,44.0,44,44,44,44.0,false,'44','44')
+sql insert into tb5 values ('2021-05-05 18:19:20',51,51.0,51,51,51,51.0,true ,'51','51')
+sql insert into tb5 values ('2021-05-05 18:19:21',52,52.0,52,52,52,52.0,true ,'52','52')
+sql insert into tb5 values ('2021-05-05 18:19:22',53,53.0,53,53,53,53.0,false,'53','53')
+sql insert into tb5 values ('2021-05-05 18:19:23',54,54.0,54,54,54,54.0,false,'54','54')
+sql insert into tb6 values ('2021-05-05 18:19:24',61,61.0,61,61,61,61.0,true ,'61','61')
+sql insert into tb6 values ('2021-05-05 18:19:25',62,62.0,62,62,62,62.0,true ,'62','62')
+sql insert into tb6 values ('2021-05-05 18:19:26',63,63.0,63,63,63,63.0,false,'63','63')
+sql insert into tb6 values ('2021-05-05 18:19:27',64,64.0,64,64,64,64.0,false,'64','64')
+sql insert into tb6 values ('2021-05-05 18:19:28',NULL,NULL,NULL,NULL,NULL,NULL,NULL,NULL,NULL)
+
+sql create table stb2 (ts timestamp, u1 int unsigned, u2 bigint unsigned, u3 smallint unsigned, u4 tinyint unsigned, ts2 timestamp) TAGS(t1 int unsigned, t2 bigint unsigned, t3 timestamp, t4 int)
+sql create table tb2_1 using stb2 tags(1,1,'2021-05-05 18:38:38',1)
+sql create table tb2_2 using stb2 tags(2,2,'2021-05-05 18:58:58',2)
+
+sql insert into tb2_1 values ('2021-05-05 18:19:00',1,2,3,4,'2021-05-05 18:28:01')
+sql insert into tb2_1 values ('2021-05-05 18:19:01',5,6,7,8,'2021-05-05 18:28:02')
+sql insert into tb2_1 values ('2021-05-05 18:19:02',2,2,3,4,'2021-05-05 18:28:03')
+sql insert into tb2_1 values ('2021-05-05 18:19:03',5,6,7,8,'2021-05-05 18:28:04')
+sql insert into tb2_1 values ('2021-05-05 18:19:04',3,2,3,4,'2021-05-05 18:28:05')
+sql insert into tb2_1 values ('2021-05-05 18:19:05',5,6,7,8,'2021-05-05 18:28:06')
+sql insert into tb2_1 values ('2021-05-05 18:19:06',4,2,3,4,'2021-05-05 18:28:07')
+sql insert into tb2_1 values ('2021-05-05 18:19:07',5,6,7,8,'2021-05-05 18:28:08')
+sql insert into tb2_1 values ('2021-05-05 18:19:08',5,2,3,4,'2021-05-05 18:28:09')
+sql insert into tb2_1 values ('2021-05-05 18:19:09',5,6,7,8,'2021-05-05 18:28:10')
+sql insert into tb2_1 values ('2021-05-05 18:19:10',6,2,3,4,'2021-05-05 18:28:11')
+sql insert into tb2_2 values ('2021-05-05 18:19:11',5,6,7,8,'2021-05-05 18:28:12')
+sql insert into tb2_2 values ('2021-05-05 18:19:12',7,2,3,4,'2021-05-05 18:28:13')
+sql insert into tb2_2 values ('2021-05-05 18:19:13',5,6,7,8,'2021-05-05 18:28:14')
+sql insert into tb2_2 values ('2021-05-05 18:19:14',8,2,3,4,'2021-05-05 18:28:15')
+sql insert into tb2_2 values ('2021-05-05 18:19:15',5,6,7,8,'2021-05-05 18:28:16')
+
+sql create table stb3 (ts timestamp, c1 int, c2 float, c3 bigint, c4 smallint, c5 tinyint, c6 double, c7 bool, c8 binary(10), c9 nchar(9)) TAGS(t1 int, t2 binary(10), t3 double)
+sql create table tb3_1 using stb3 tags(1,'1',1.0)
+sql create table tb3_2 using stb3 tags(2,'2',2.0)
+
+sql insert into tb3_1 values ('2021-01-05 18:19:00',1,1.0,1,1,1,1.0,true ,'1','1')
+sql insert into tb3_1 values ('2021-02-05 18:19:01',2,2.0,2,2,2,2.0,true ,'2','2')
+sql insert into tb3_1 values ('2021-03-05 18:19:02',3,3.0,3,3,3,3.0,false,'3','3')
+sql insert into tb3_1 values ('2021-04-05 18:19:03',4,4.0,4,4,4,4.0,false,'4','4')
+sql insert into tb3_1 values ('2021-05-05 18:19:28',5,NULL,5,NULL,5,NULL,true,NULL,'5')
+sql insert into tb3_1 values ('2021-06-05 18:19:28',NULL,6.0,NULL,6,NULL,6.0,NULL,'6',NULL)
+sql insert into tb3_1 values ('2021-07-05 18:19:28',NULL,NULL,NULL,NULL,NULL,NULL,NULL,NULL,NULL)
+sql insert into tb3_2 values ('2021-01-06 18:19:00',11,11.0,11,11,11,11.0,true ,'11','11')
+sql insert into tb3_2 values ('2021-02-06 18:19:01',12,12.0,12,12,12,12.0,true ,'12','12')
+sql insert into tb3_2 values ('2021-03-06 18:19:02',13,13.0,13,13,13,13.0,false,'13','13')
+sql insert into tb3_2 values ('2021-04-06 18:19:03',14,14.0,14,14,14,14.0,false,'14','14')
+sql insert into tb3_2 values ('2021-05-06 18:19:28',15,NULL,15,NULL,15,NULL,true,NULL,'15')
+sql insert into tb3_2 values ('2021-06-06 18:19:28',NULL,16.0,NULL,16,NULL,16.0,NULL,'16',NULL)
+sql insert into tb3_2 values ('2021-07-06 18:19:28',NULL,NULL,NULL,NULL,NULL,NULL,NULL,NULL,NULL)
+
+sql create table stb4 (ts timestamp, c1 int, c2 float, c3 bigint, c4 smallint, c5 tinyint, c6 double, c7 bool, c8 binary(10), c9 nchar(9),c10 binary(16300)) TAGS(t1 int, t2 binary(10), t3 double)
+sql create table tb4_0 using stb4 tags(0,'0',0.0)
+sql create table tb4_1 using stb4 tags(1,'1',1.0)
+sql create table tb4_2 using stb4 tags(2,'2',2.0)
+sql create table tb4_3 using stb4 tags(3,'3',3.0)
+sql create table tb4_4 using stb4 tags(4,'4',4.0)
+
+$i = 0
+$ts0 = 1625850000000
+$blockNum = 5
+$delta = 0
+$tbname0 = tb4_
+$a = 0
+$b = 200
+$c = 400
+while $i < $blockNum
+  $x = 0
+  $rowNum = 1200
+  while $x < $rowNum
+    $ts = $ts0 + $x
+    $a = $a + 1
+    $b = $b + 1
+    $c = $c + 1
+    $d = $x / 10
+    $tin = $rowNum
+    $binary = 'binary . $c
+    $binary = $binary . '
+    $nchar = 'nchar . $c
+    $nchar = $nchar . '
+    $tbname = 'tb4_ . $i
+    $tbname = $tbname . '
+    sql insert into $tbname values ( $ts , $a , $b , $c , $d , $d , $c , true, $binary , $nchar , $binary )
+    $x = $x + 1
+  endw
+
+  $i = $i + 1
+  $ts0 = $ts0 + 259200000
+endw
+
+run tsim/parser/condition_query.sim
+
+print ================== restart server to commit data into disk
+system sh/exec.sh -n dnode1 -s stop -x SIGINT
+system sh/exec.sh -n dnode1 -s start
+print ================== server restart completed
+sql connect
+
+run tsim/parser/condition_query.sim
+

tests/script/tsim/query/emptyTsRange_scl.sim

+system sh/stop_dnodes.sh
+system sh/deploy.sh -n dnode1 -i 1
+system sh/cfg.sh -n dnode1 -c filterScalarMode -v 1
+system sh/exec.sh -n dnode1 -s start
+sql connect
+
+sql drop database if exists db1;
+sql create database if not exists db1;
+sql use db1;
+sql create stable sta (ts timestamp, f1 double, f2 binary(200)) tags(t1 int);
+sql create table tba1 using sta tags(1);
+sql insert into tba1 values ('2022-04-26 15:15:01', 1.0, "a");
+sql insert into tba1 values ('2022-04-26 15:15:02', 2.0, "b");
+sql insert into tba1 values ('2022-04-26 15:15:04', 4.0, "b");
+sql insert into tba1 values ('2022-04-26 15:15:05', 5.0, "b");
+sql select last_row(*) from sta where ts >= 1678901803783 and ts <= 1678901803783 and  _c0 <= 1678901803782 interval(10d,8d) fill(linear) order by _wstart desc;
+if $rows != 0 then
+  return -1
+endi
+
+system sh/exec.sh -n dnode1 -s stop -x SIGINT