Merge pull request #11642 from taosdata/fix/TD-14843

refactor(query): refactor cast function

include/util/taoserror.h

@@ -617,6 +617,7 @@ int32_t* taosGetErrno();
 #define TSDB_CODE_FUNC_FUNTION_ERROR            TAOS_DEF_ERROR_CODE(0, 0x2800)
 #define TSDB_CODE_FUNC_FUNTION_PARA_NUM         TAOS_DEF_ERROR_CODE(0, 0x2801)
 #define TSDB_CODE_FUNC_FUNTION_PARA_TYPE        TAOS_DEF_ERROR_CODE(0, 0x2802)
+#define TSDB_CODE_FUNC_FUNTION_PARA_VALUE       TAOS_DEF_ERROR_CODE(0, 0x2803)
 
 #ifdef __cplusplus
 }

source/libs/function/src/builtins.c

@@ -28,11 +28,15 @@ static int32_t buildFuncErrMsg(char* pErrBuf, int32_t len, int32_t errCode, cons
 }
 
 static int32_t invaildFuncParaNumErrMsg(char* pErrBuf, int32_t len, const char* pFuncName) {
-  return buildFuncErrMsg(pErrBuf, len, TSDB_CODE_FUNC_FUNTION_PARA_NUM, "Invalid number of arguments : %s", pFuncName);
+  return buildFuncErrMsg(pErrBuf, len, TSDB_CODE_FUNC_FUNTION_PARA_NUM, "Invalid number of parameters : %s", pFuncName);
 }
 
 static int32_t invaildFuncParaTypeErrMsg(char* pErrBuf, int32_t len, const char* pFuncName) {
-  return buildFuncErrMsg(pErrBuf, len, TSDB_CODE_FUNC_FUNTION_PARA_TYPE, "Inconsistent datatypes : %s", pFuncName);
+  return buildFuncErrMsg(pErrBuf, len, TSDB_CODE_FUNC_FUNTION_PARA_TYPE, "Invalid datatypes : %s", pFuncName);
+}
+
+static int32_t invaildFuncParaValueErrMsg(char* pErrBuf, int32_t len, const char* pFuncName) {
+  return buildFuncErrMsg(pErrBuf, len, TSDB_CODE_FUNC_FUNTION_PARA_VALUE, "Invalid parameter value : %s", pFuncName);
 }
 
 // There is only one parameter of numeric type, and the return type is parameter type
@@ -318,10 +322,15 @@ static int32_t translateCast(SFunctionNode* pFunc, char* pErrBuf, int32_t len) {
   uint8_t para1Type = ((SExprNode*)nodesListGetNode(pFunc->pParameterList, 0))->resType.type;
   // The function return type has been set during syntax parsing
   uint8_t para2Type = pFunc->node.resType.type;
-  if ((TSDB_DATA_TYPE_JSON == para1Type || TSDB_DATA_TYPE_BLOB == para1Type || TSDB_DATA_TYPE_MEDIUMBLOB == para1Type) ||
-      (TSDB_DATA_TYPE_JSON == para2Type || TSDB_DATA_TYPE_BLOB == para2Type || TSDB_DATA_TYPE_MEDIUMBLOB == para2Type)) {
+  if (para2Type != TSDB_DATA_TYPE_BIGINT && para2Type != TSDB_DATA_TYPE_UBIGINT &&
+      para2Type != TSDB_DATA_TYPE_VARCHAR && para2Type != TSDB_DATA_TYPE_NCHAR &&
+      para2Type != TSDB_DATA_TYPE_TIMESTAMP) {
     return invaildFuncParaTypeErrMsg(pErrBuf, len, pFunc->functionName);
   }
+  int32_t para2Bytes = pFunc->node.resType.bytes;
+  if (para2Bytes <= 0) { //non-positive value or overflow
+    return invaildFuncParaValueErrMsg(pErrBuf, len, pFunc->functionName);
+  }
   return TSDB_CODE_SUCCESS;
 }
 

source/libs/scalar/src/sclfunc.c

@@ -659,27 +659,22 @@ int32_t substrFunction(SScalarParam *pInput, int32_t inputNum, SScalarParam *pOu
 int32_t castFunction(SScalarParam *pInput, int32_t inputNum, SScalarParam *pOutput) {
   int16_t inputType  = pInput[0].columnData->info.type;
   int16_t outputType = pOutput[0].columnData->info.type;
-  if (outputType != TSDB_DATA_TYPE_BIGINT && outputType != TSDB_DATA_TYPE_UBIGINT &&
-      outputType != TSDB_DATA_TYPE_VARCHAR && outputType != TSDB_DATA_TYPE_NCHAR &&
-      outputType != TSDB_DATA_TYPE_TIMESTAMP) {
-    return TSDB_CODE_FAILED;
-  }
   int64_t outputLen = pOutput[0].columnData->info.bytes;
 
-  char *input = NULL;
+  if (IS_VAR_DATA_TYPE(outputType)) {
+    int32_t factor = (TSDB_DATA_TYPE_NCHAR == outputType) ? TSDB_NCHAR_SIZE : 1;
+    outputLen = outputLen * factor + VARSTR_HEADER_SIZE;
+  }
+
   char *outputBuf = taosMemoryCalloc(outputLen * pInput[0].numOfRows, 1);
   char *output = outputBuf;
-  if (IS_VAR_DATA_TYPE(inputType)) {
-    input = pInput[0].columnData->pData + pInput[0].columnData->varmeta.offset[0];
-  } else {
-    input = pInput[0].columnData->pData;
-  }
 
   for (int32_t i = 0; i < pInput[0].numOfRows; ++i) {
     if (colDataIsNull_s(pInput[0].columnData, i)) {
       colDataAppendNULL(pOutput->columnData, i);
       continue;
     }
+    char *input = colDataGetData(pInput[0].columnData, i);
 
     switch(outputType) {
       case TSDB_DATA_TYPE_BIGINT: {
@@ -736,7 +731,7 @@ int32_t castFunction(SScalarParam *pInput, int32_t inputNum, SScalarParam *pOutp
         } else if (inputType == TSDB_DATA_TYPE_BINARY) {
           int32_t len = sprintf(varDataVal(output), "%.*s", (int32_t)(outputLen - VARSTR_HEADER_SIZE), varDataVal(input));
           varDataSetLen(output, len);
-        } else if (inputType == TSDB_DATA_TYPE_BINARY || inputType == TSDB_DATA_TYPE_NCHAR) {
+        } else if (inputType == TSDB_DATA_TYPE_NCHAR) {
           //not support
           return TSDB_CODE_FAILED;
         } else {
@@ -789,11 +784,6 @@ int32_t castFunction(SScalarParam *pInput, int32_t inputNum, SScalarParam *pOutp
     }
 
     colDataAppend(pOutput->columnData, i, output, false);
-    if (IS_VAR_DATA_TYPE(inputType)) {
-      input  += varDataTLen(input);
-    } else {
-      input  += tDataTypes[inputType].bytes;
-    }
     if (IS_VAR_DATA_TYPE(outputType)) {
       output += varDataTLen(output);
     } else {