Merge pull request #16179 from taosdata/fix/TD-17499

fix(query): set proper ltrim/rtrim result schema bytes

source/libs/function/src/builtins.c

@@ -192,6 +192,24 @@ static bool validateTimezoneFormat(const SValueNode* pVal) {
   return true;
 }
 
+static int32_t countTrailingSpaces(const SValueNode* pVal, bool isLtrim) {
+  int32_t numOfSpaces = 0;
+  int32_t len = varDataLen(pVal->datum.p);
+  char*   str = varDataVal(pVal->datum.p);
+
+  int32_t startPos = isLtrim ? 0 : len - 1;
+  int32_t step = isLtrim ? 1 : -1;
+  for (int32_t i = startPos; i < len || i >= 0; i += step) {
+    if (!isspace(str[i])) {
+      break;
+    }
+    numOfSpaces++;
+  }
+
+  return numOfSpaces;
+
+}
+
 void static addTimezoneParam(SNodeList* pList) {
   char      buf[6] = {0};
   time_t    t = taosTime(NULL);
@@ -293,6 +311,40 @@ static int32_t translateInOutStr(SFunctionNode* pFunc, char* pErrBuf, int32_t le
   return TSDB_CODE_SUCCESS;
 }
 
+static int32_t translateTrimStr(SFunctionNode* pFunc, char* pErrBuf, int32_t len, bool isLtrim) {
+  if (1 != LIST_LENGTH(pFunc->pParameterList)) {
+    return invaildFuncParaNumErrMsg(pErrBuf, len, pFunc->functionName);
+  }
+
+  SExprNode* pPara1 = (SExprNode*)nodesListGetNode(pFunc->pParameterList, 0);
+  if (!IS_VAR_DATA_TYPE(pPara1->resType.type)) {
+    return invaildFuncParaTypeErrMsg(pErrBuf, len, pFunc->functionName);
+  }
+
+  int32_t numOfSpaces = 0;
+  SNode* pParamNode1 = nodesListGetNode(pFunc->pParameterList, 0);
+  // for select trim functions with constant value from table,
+  // need to set the proper result result schema bytes to avoid
+  // trailing garbage characters
+  if (nodeType(pParamNode1) == QUERY_NODE_VALUE) {
+    SValueNode* pValue = (SValueNode*)pParamNode1;
+    numOfSpaces = countTrailingSpaces(pValue, isLtrim);
+  }
+
+
+  int32_t resBytes = pPara1->resType.bytes - numOfSpaces;
+  pFunc->node.resType = (SDataType){.bytes = resBytes, .type = pPara1->resType.type};
+  return TSDB_CODE_SUCCESS;
+}
+
+static int32_t translateLtrim(SFunctionNode* pFunc, char* pErrBuf, int32_t len) {
+  return translateTrimStr(pFunc, pErrBuf, len, true);
+}
+
+static int32_t translateRtrim(SFunctionNode* pFunc, char* pErrBuf, int32_t len) {
+  return translateTrimStr(pFunc, pErrBuf, len, false);
+}
+
 static int32_t translateLogarithm(SFunctionNode* pFunc, char* pErrBuf, int32_t len) {
   int32_t numOfParams = LIST_LENGTH(pFunc->pParameterList);
   if (1 != numOfParams && 2 != numOfParams) {
@@ -2827,7 +2879,7 @@ const SBuiltinFuncDefinition funcMgtBuiltins[] = {
     .name = "ltrim",
     .type = FUNCTION_TYPE_LTRIM,
     .classification = FUNC_MGT_SCALAR_FUNC | FUNC_MGT_STRING_FUNC,
-    .translateFunc = translateInOutStr,
+    .translateFunc = translateLtrim,
     .getEnvFunc   = NULL,
     .initFunc     = NULL,
     .sprocessFunc = ltrimFunction,
@@ -2837,7 +2889,7 @@ const SBuiltinFuncDefinition funcMgtBuiltins[] = {
     .name = "rtrim",
     .type = FUNCTION_TYPE_RTRIM,
     .classification = FUNC_MGT_SCALAR_FUNC | FUNC_MGT_STRING_FUNC,
-    .translateFunc = translateInOutStr,
+    .translateFunc = translateRtrim,
     .getEnvFunc   = NULL,
     .initFunc     = NULL,
     .sprocessFunc = rtrimFunction,

source/libs/scalar/src/scalar.c

@@ -758,7 +758,9 @@ EDealRes sclRewriteFunction(SNode** pNode, SScalarCtx *ctx) {
       res->datum.p = taosMemoryCalloc(len, 1);
       memcpy(res->datum.p, output.columnData->pData, len);
     } else if (IS_VAR_DATA_TYPE(type)) {
-      res->datum.p = taosMemoryCalloc(res->node.resType.bytes + VARSTR_HEADER_SIZE + 1, 1);
+      //res->datum.p = taosMemoryCalloc(res->node.resType.bytes + VARSTR_HEADER_SIZE + 1, 1);
+      res->datum.p = taosMemoryCalloc(varDataTLen(output.columnData->pData), 1);
+      res->node.resType.bytes = varDataTLen(output.columnData->pData);
       memcpy(res->datum.p, output.columnData->pData, varDataTLen(output.columnData->pData));
     } else {
       nodesSetValueNodeValue(res, output.columnData->pData);