tscParseLineProtocol.c 70.7 KB
Newer Older
S
shenglian zhou 已提交
1 2 3 4
#include <ctype.h>
#include <stdio.h>
#include <stdlib.h>
#include <string.h>
5

S
shenglian zhou 已提交
6 7 8 9 10 11 12 13
#include "os.h"
#include "osString.h"
#include "ttype.h"
#include "tmd5.h"
#include "tstrbuild.h"
#include "tname.h"
#include "hash.h"
#include "tskiplist.h"
14

S
shenglian zhou 已提交
15
#include "tscUtil.h"
16 17
#include "tsclient.h"
#include "tscLog.h"
S
shenglian zhou 已提交
18

19
#include "taos.h"
20
#include "tscParseLine.h"
21

S
shenglian zhou 已提交
22 23 24 25 26 27
typedef struct  {
  char sTableName[TSDB_TABLE_NAME_LEN];
  SHashObj* tagHash;
  SHashObj* fieldHash;
  SArray* tags; //SArray<SSchema>
  SArray* fields; //SArray<SSchema>
S
shenglian zhou 已提交
28
  uint8_t precision;
S
shenglian zhou 已提交
29 30
} SSmlSTableSchema;

31 32
//=================================================================================================

33 34 35 36 37 38 39 40 41 42 43 44
static uint64_t linesSmlHandleId = 0;

uint64_t genLinesSmlId() {
  uint64_t id;

  do {
    id = atomic_add_fetch_64(&linesSmlHandleId, 1);
  } while (id == 0);

  return id;
}

S
shenglian zhou 已提交
45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84
int compareSmlColKv(const void* p1, const void* p2) {
  TAOS_SML_KV* kv1 = (TAOS_SML_KV*)p1;
  TAOS_SML_KV* kv2 = (TAOS_SML_KV*)p2;
  int kvLen1 = (int)strlen(kv1->key);
  int kvLen2 = (int)strlen(kv2->key);
  int res = strncasecmp(kv1->key, kv2->key, MIN(kvLen1, kvLen2));
  if (res != 0) {
    return res;
  } else {
    return kvLen1-kvLen2;
  }
}

typedef enum {
  SCHEMA_ACTION_CREATE_STABLE,
  SCHEMA_ACTION_ADD_COLUMN,
  SCHEMA_ACTION_ADD_TAG,
  SCHEMA_ACTION_CHANGE_COLUMN_SIZE,
  SCHEMA_ACTION_CHANGE_TAG_SIZE,
} ESchemaAction;

typedef struct {
  char sTableName[TSDB_TABLE_NAME_LEN];
  SArray* tags; //SArray<SSchema>
  SArray* fields; //SArray<SSchema>
} SCreateSTableActionInfo;

typedef struct {
  char sTableName[TSDB_TABLE_NAME_LEN];
  SSchema* field;
} SAlterSTableActionInfo;

typedef struct {
  ESchemaAction action;
  union {
    SCreateSTableActionInfo createSTable;
    SAlterSTableActionInfo alterSTable;
  };
} SSchemaAction;

S
shenglian zhou 已提交
85
static int32_t getFieldBytesFromSmlKv(TAOS_SML_KV* kv, int32_t* bytes, uint64_t id) {
S
shenglian zhou 已提交
86 87 88 89 90 91
  if (!IS_VAR_DATA_TYPE(kv->type)) {
    *bytes = tDataTypes[kv->type].bytes;
  } else {
    if (kv->type == TSDB_DATA_TYPE_NCHAR) {
      char* ucs = malloc(kv->length * TSDB_NCHAR_SIZE + 1);
      int32_t bytesNeeded = 0;
S
Shenglian Zhou 已提交
92 93 94
      bool succ = taosMbsToUcs4(kv->value, kv->length, ucs, kv->length * TSDB_NCHAR_SIZE, &bytesNeeded);
      if (!succ) {
        free(ucs);
S
shenglian zhou 已提交
95
        tscError("SML:0x%"PRIx64" convert nchar string to UCS4_LE failed:%s", id, kv->value);
S
Shenglian Zhou 已提交
96 97
        return TSDB_CODE_TSC_INVALID_VALUE;
      }
S
shenglian zhou 已提交
98 99 100 101 102 103 104 105 106
      free(ucs);
      *bytes =  bytesNeeded + VARSTR_HEADER_SIZE;
    } else if (kv->type == TSDB_DATA_TYPE_BINARY) {
      *bytes = kv->length + VARSTR_HEADER_SIZE;
    }
  }
  return 0;
}

S
shenglian zhou 已提交
107
static int32_t buildSmlKvSchema(TAOS_SML_KV* smlKv, SHashObj* hash, SArray* array, SSmlLinesInfo* info) {
S
shenglian zhou 已提交
108
  SSchema* pField = NULL;
109 110
  size_t* pFieldIdx = taosHashGet(hash, smlKv->key, strlen(smlKv->key));
  size_t fieldIdx = -1;
S
Shenglian Zhou 已提交
111
  int32_t code = 0;
112 113 114
  if (pFieldIdx) {
    fieldIdx = *pFieldIdx;
    pField = taosArrayGet(array, fieldIdx);
S
shenglian zhou 已提交
115 116

    if (pField->type != smlKv->type) {
S
shenglian zhou 已提交
117
      tscError("SML:0x%"PRIx64" type mismatch. key %s, type %d. type before %d", info->id, smlKv->key, smlKv->type, pField->type);
S
Shenglian Zhou 已提交
118
      return TSDB_CODE_TSC_INVALID_VALUE;
S
shenglian zhou 已提交
119 120 121
    }

    int32_t bytes = 0;
S
shenglian zhou 已提交
122
    code = getFieldBytesFromSmlKv(smlKv, &bytes, info->id);
S
Shenglian Zhou 已提交
123 124 125
    if (code != 0) {
      return code;
    }
S
shenglian zhou 已提交
126 127 128
    pField->bytes = MAX(pField->bytes, bytes);

  } else {
S
shenglian zhou 已提交
129
    SSchema field = {0};
S
shenglian zhou 已提交
130 131 132 133 134 135
    size_t tagKeyLen = strlen(smlKv->key);
    strncpy(field.name, smlKv->key, tagKeyLen);
    field.name[tagKeyLen] = '\0';
    field.type = smlKv->type;

    int32_t bytes = 0;
S
shenglian zhou 已提交
136
    code = getFieldBytesFromSmlKv(smlKv, &bytes, info->id);
S
Shenglian Zhou 已提交
137 138 139
    if (code != 0) {
      return code;
    }
S
shenglian zhou 已提交
140 141 142
    field.bytes = bytes;

    pField = taosArrayPush(array, &field);
143 144
    fieldIdx = taosArrayGetSize(array) - 1;
    taosHashPut(hash, field.name, tagKeyLen, &fieldIdx, sizeof(fieldIdx));
S
shenglian zhou 已提交
145
  }
146

147 148
  uintptr_t valPointer = (uintptr_t)smlKv;
  taosHashPut(info->smlDataToSchema, &valPointer, sizeof(uintptr_t), &fieldIdx, sizeof(fieldIdx));
149

S
shenglian zhou 已提交
150 151 152
  return 0;
}

153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186
static int32_t getSmlMd5ChildTableName(TAOS_SML_DATA_POINT* point, char* tableName, int* tableNameLen,
                                       SSmlLinesInfo* info) {
  tscDebug("SML:0x%"PRIx64" taos_sml_insert get child table name through md5", info->id);
  qsort(point->tags, point->tagNum, sizeof(TAOS_SML_KV), compareSmlColKv);

  SStringBuilder sb; memset(&sb, 0, sizeof(sb));
  char sTableName[TSDB_TABLE_NAME_LEN] = {0};
  strtolower(sTableName, point->stableName);
  taosStringBuilderAppendString(&sb, sTableName);
  for (int j = 0; j < point->tagNum; ++j) {
    taosStringBuilderAppendChar(&sb, ',');
    TAOS_SML_KV* tagKv = point->tags + j;
    char tagName[TSDB_COL_NAME_LEN] = {0};
    strtolower(tagName, tagKv->key);
    taosStringBuilderAppendString(&sb, tagName);
    taosStringBuilderAppendChar(&sb, '=');
    taosStringBuilderAppend(&sb, tagKv->value, tagKv->length);
  }
  size_t len = 0;
  char* keyJoined = taosStringBuilderGetResult(&sb, &len);
  MD5_CTX context;
  MD5Init(&context);
  MD5Update(&context, (uint8_t *)keyJoined, (uint32_t)len);
  MD5Final(&context);
  *tableNameLen = snprintf(tableName, *tableNameLen,
                           "t_%02x%02x%02x%02x%02x%02x%02x%02x%02x%02x%02x%02x%02x%02x%02x%02x", context.digest[0],
                           context.digest[1], context.digest[2], context.digest[3], context.digest[4], context.digest[5], context.digest[6],
                           context.digest[7], context.digest[8], context.digest[9], context.digest[10], context.digest[11],
                           context.digest[12], context.digest[13], context.digest[14], context.digest[15]);
  taosStringBuilderDestroy(&sb);
  tscDebug("SML:0x%"PRIx64" child table name: %s", info->id, tableName);
  return 0;
}

S
shenglian zhou 已提交
187
static int32_t buildDataPointSchemas(TAOS_SML_DATA_POINT* points, int numPoint, SArray* stableSchemas, SSmlLinesInfo* info) {
S
Shenglian Zhou 已提交
188
  int32_t code = 0;
S
shenglian zhou 已提交
189 190 191 192 193 194
  SHashObj* sname2shema = taosHashInit(32,
                                       taosGetDefaultHashFunction(TSDB_DATA_TYPE_BINARY), true, false);

  for (int i = 0; i < numPoint; ++i) {
    TAOS_SML_DATA_POINT* point = &points[i];
    size_t stableNameLen = strlen(point->stableName);
195
    size_t* pStableIdx = taosHashGet(sname2shema, point->stableName, stableNameLen);
S
shenglian zhou 已提交
196
    SSmlSTableSchema* pStableSchema = NULL;
197 198 199 200
    size_t stableIdx = -1;
    if (pStableIdx) {
      pStableSchema= taosArrayGet(stableSchemas, *pStableIdx);
      stableIdx = *pStableIdx;
S
shenglian zhou 已提交
201 202 203 204 205 206 207 208 209 210
    } else {
      SSmlSTableSchema schema;
      strncpy(schema.sTableName, point->stableName, stableNameLen);
      schema.sTableName[stableNameLen] = '\0';
      schema.fields = taosArrayInit(64, sizeof(SSchema));
      schema.tags = taosArrayInit(8, sizeof(SSchema));
      schema.tagHash = taosHashInit(8, taosGetDefaultHashFunction(TSDB_DATA_TYPE_BINARY), true, false);
      schema.fieldHash = taosHashInit(64, taosGetDefaultHashFunction(TSDB_DATA_TYPE_BINARY), true, false);

      pStableSchema = taosArrayPush(stableSchemas, &schema);
211 212
      stableIdx = taosArrayGetSize(stableSchemas) - 1;
      taosHashPut(sname2shema, schema.sTableName, stableNameLen, &stableIdx, sizeof(size_t));
S
shenglian zhou 已提交
213 214 215 216
    }

    for (int j = 0; j < point->tagNum; ++j) {
      TAOS_SML_KV* tagKv = point->tags + j;
217 218 219 220 221 222 223 224 225
      if (!point->childTableName) {
        char childTableName[TSDB_TABLE_NAME_LEN];
        int32_t tableNameLen = TSDB_TABLE_NAME_LEN;
        getSmlMd5ChildTableName(point, childTableName, &tableNameLen, info);
        point->childTableName = calloc(1, tableNameLen+1);
        strncpy(point->childTableName, childTableName, tableNameLen);
        point->childTableName[tableNameLen] = '\0';
      }

S
shenglian zhou 已提交
226
      code = buildSmlKvSchema(tagKv, pStableSchema->tagHash, pStableSchema->tags, info);
S
Shenglian Zhou 已提交
227
      if (code != 0) {
S
shenglian zhou 已提交
228
        tscError("SML:0x%"PRIx64" build data point schema failed. point no.: %d, tag key: %s", info->id, i, tagKv->key);
S
Shenglian Zhou 已提交
229 230
        return code;
      }
S
shenglian zhou 已提交
231 232 233 234
    }

    for (int j = 0; j < point->fieldNum; ++j) {
      TAOS_SML_KV* fieldKv = point->fields + j;
S
shenglian zhou 已提交
235
      code = buildSmlKvSchema(fieldKv, pStableSchema->fieldHash, pStableSchema->fields, info);
S
Shenglian Zhou 已提交
236
      if (code != 0) {
S
shenglian zhou 已提交
237
        tscError("SML:0x%"PRIx64" build data point schema failed. point no.: %d, tag key: %s", info->id, i, fieldKv->key);
S
Shenglian Zhou 已提交
238 239
        return code;
      }
S
shenglian zhou 已提交
240 241
    }

242 243
    uintptr_t valPointer = (uintptr_t)point;
    taosHashPut(info->smlDataToSchema, &valPointer, sizeof(uintptr_t), &stableIdx, sizeof(stableIdx));
S
shenglian zhou 已提交
244 245 246 247 248 249 250 251 252 253
  }

  size_t numStables = taosArrayGetSize(stableSchemas);
  for (int32_t i = 0; i < numStables; ++i) {
    SSmlSTableSchema* schema = taosArrayGet(stableSchemas, i);
    taosHashCleanup(schema->tagHash);
    taosHashCleanup(schema->fieldHash);
  }
  taosHashCleanup(sname2shema);

S
shenglian zhou 已提交
254
  tscDebug("SML:0x%"PRIx64" build point schema succeed. num of super table: %zu", info->id, numStables);
S
Shenglian Zhou 已提交
255 256 257 258 259 260
  for (int32_t i = 0; i < numStables; ++i) {
    SSmlSTableSchema* schema = taosArrayGet(stableSchemas, i);
    tscDebug("\ttable name: %s, tags number: %zu, fields number: %zu", schema->sTableName,
             taosArrayGetSize(schema->tags), taosArrayGetSize(schema->fields));
  }

S
shenglian zhou 已提交
261 262 263
  return 0;
}

264
static int32_t generateSchemaAction(SSchema* pointColField, SHashObj* dbAttrHash, SArray* dbAttrArray, bool isTag, char sTableName[],
S
shenglian zhou 已提交
265
                                       SSchemaAction* action, bool* actionNeeded, SSmlLinesInfo* info) {
266 267 268 269
  char fieldNameLowerCase[TSDB_COL_NAME_LEN] = {0};
  strtolower(fieldNameLowerCase, pointColField->name);

  size_t* pDbIndex = taosHashGet(dbAttrHash, fieldNameLowerCase, strlen(fieldNameLowerCase));
270 271 272
  if (pDbIndex) {
    SSchema* dbAttr = taosArrayGet(dbAttrArray, *pDbIndex);
    assert(strcasecmp(dbAttr->name, pointColField->name) == 0);
S
shenglian zhou 已提交
273
    if (pointColField->type != dbAttr->type) {
S
shenglian zhou 已提交
274
      tscError("SML:0x%"PRIx64" point type and db type mismatch. key: %s. point type: %d, db type: %d", info->id, pointColField->name,
S
Shenglian Zhou 已提交
275 276
               pointColField->type, dbAttr->type);
      return TSDB_CODE_TSC_INVALID_VALUE;
S
shenglian zhou 已提交
277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300
    }

    if (IS_VAR_DATA_TYPE(pointColField->type) && (pointColField->bytes > dbAttr->bytes)) {
      if (isTag) {
        action->action = SCHEMA_ACTION_CHANGE_TAG_SIZE;
      } else {
        action->action = SCHEMA_ACTION_CHANGE_COLUMN_SIZE;
      }
      memset(&action->alterSTable, 0,  sizeof(SAlterSTableActionInfo));
      memcpy(action->alterSTable.sTableName, sTableName, TSDB_TABLE_NAME_LEN);
      action->alterSTable.field = pointColField;
      *actionNeeded = true;
    }
  } else {
    if (isTag) {
      action->action = SCHEMA_ACTION_ADD_TAG;
    } else {
      action->action = SCHEMA_ACTION_ADD_COLUMN;
    }
    memset(&action->alterSTable, 0, sizeof(SAlterSTableActionInfo));
    memcpy(action->alterSTable.sTableName, sTableName, TSDB_TABLE_NAME_LEN);
    action->alterSTable.field = pointColField;
    *actionNeeded = true;
  }
S
shenglian zhou 已提交
301 302 303 304
  if (*actionNeeded) {
    tscDebug("SML:0x%" PRIx64 " generate schema action. column name: %s, action: %d", info->id, fieldNameLowerCase,
             action->action);
  }
S
shenglian zhou 已提交
305 306 307
  return 0;
}

S
shenglian zhou 已提交
308
static int32_t buildColumnDescription(SSchema* field,
S
shenglian zhou 已提交
309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328
                               char* buf, int32_t bufSize, int32_t* outBytes) {
  uint8_t type = field->type;

  if (type == TSDB_DATA_TYPE_BINARY || type == TSDB_DATA_TYPE_NCHAR) {
    int32_t bytes = field->bytes - VARSTR_HEADER_SIZE;
    if (type == TSDB_DATA_TYPE_NCHAR) {
      bytes =  bytes/TSDB_NCHAR_SIZE;
    }
    int out = snprintf(buf, bufSize,"%s %s(%d)",
                       field->name,tDataTypes[field->type].name, bytes);
    *outBytes = out;
  } else {
    int out = snprintf(buf, bufSize, "%s %s",
                       field->name, tDataTypes[type].name);
    *outBytes = out;
  }

  return 0;
}

S
shenglian zhou 已提交
329

S
shenglian zhou 已提交
330
static int32_t applySchemaAction(TAOS* taos, SSchemaAction* action, SSmlLinesInfo* info) {
S
shenglian zhou 已提交
331 332
  int32_t code = 0;
  int32_t outBytes = 0;
333 334
  char *result = (char *)calloc(1, tsMaxSQLStringLen+1);
  int32_t capacity = tsMaxSQLStringLen +  1;
S
shenglian zhou 已提交
335

S
shenglian zhou 已提交
336
  tscDebug("SML:0x%"PRIx64" apply schema action. action: %d", info->id, action->action);
S
shenglian zhou 已提交
337 338 339 340 341 342
  switch (action->action) {
    case SCHEMA_ACTION_ADD_COLUMN: {
      int n = sprintf(result, "alter stable %s add column ", action->alterSTable.sTableName);
      buildColumnDescription(action->alterSTable.field, result+n, capacity-n, &outBytes);
      TAOS_RES* res = taos_query(taos, result); //TODO async doAsyncQuery
      code = taos_errno(res);
S
shenglian zhou 已提交
343 344 345
      char* errStr = taos_errstr(res);
      char* begin = strstr(errStr, "duplicated column names");
      bool tscDupColNames = (begin != NULL);
S
shenglian zhou 已提交
346
      if (code != TSDB_CODE_SUCCESS) {
S
shenglian zhou 已提交
347
        tscError("SML:0x%"PRIx64" apply schema action. error: %s", info->id, errStr);
S
shenglian zhou 已提交
348
      }
349 350
      taos_free_result(res);

S
shenglian zhou 已提交
351
      if (code == TSDB_CODE_MND_FIELD_ALREAY_EXIST || code == TSDB_CODE_MND_TAG_ALREAY_EXIST || tscDupColNames) {
352 353
        TAOS_RES* res2 = taos_query(taos, "RESET QUERY CACHE");
        code = taos_errno(res2);
354 355 356
        if (code != TSDB_CODE_SUCCESS) {
          tscError("SML:0x%" PRIx64 " apply schema action. reset query cache. error: %s", info->id, taos_errstr(res2));
        }
357 358
        taos_free_result(res2);
      }
S
shenglian zhou 已提交
359 360 361 362 363 364 365 366
      break;
    }
    case SCHEMA_ACTION_ADD_TAG: {
      int n = sprintf(result, "alter stable %s add tag ", action->alterSTable.sTableName);
      buildColumnDescription(action->alterSTable.field,
                             result+n, capacity-n, &outBytes);
      TAOS_RES* res = taos_query(taos, result); //TODO async doAsyncQuery
      code = taos_errno(res);
S
shenglian zhou 已提交
367 368 369
      char* errStr = taos_errstr(res);
      char* begin = strstr(errStr, "duplicated column names");
      bool tscDupColNames = (begin != NULL);
S
shenglian zhou 已提交
370 371 372
      if (code != TSDB_CODE_SUCCESS) {
        tscError("SML:0x%"PRIx64" apply schema action. error : %s", info->id, taos_errstr(res));
      }
373 374
      taos_free_result(res);

S
shenglian zhou 已提交
375
      if (code == TSDB_CODE_MND_TAG_ALREAY_EXIST || code == TSDB_CODE_MND_FIELD_ALREAY_EXIST || tscDupColNames) {
376 377
        TAOS_RES* res2 = taos_query(taos, "RESET QUERY CACHE");
        code = taos_errno(res2);
378 379 380
        if (code != TSDB_CODE_SUCCESS) {
          tscError("SML:0x%" PRIx64 " apply schema action. reset query cache. error: %s", info->id, taos_errstr(res2));
        }
381 382
        taos_free_result(res2);
      }
S
shenglian zhou 已提交
383 384 385 386 387 388 389 390
      break;
    }
    case SCHEMA_ACTION_CHANGE_COLUMN_SIZE: {
      int n = sprintf(result, "alter stable %s modify column ", action->alterSTable.sTableName);
      buildColumnDescription(action->alterSTable.field, result+n,
                             capacity-n, &outBytes);
      TAOS_RES* res = taos_query(taos, result); //TODO async doAsyncQuery
      code = taos_errno(res);
S
shenglian zhou 已提交
391 392 393
      if (code != TSDB_CODE_SUCCESS) {
        tscError("SML:0x%"PRIx64" apply schema action. error : %s", info->id, taos_errstr(res));
      }
394 395
      taos_free_result(res);

396
      if (code == TSDB_CODE_MND_INVALID_COLUMN_LENGTH || code == TSDB_CODE_TSC_INVALID_COLUMN_LENGTH) {
397 398
        TAOS_RES* res2 = taos_query(taos, "RESET QUERY CACHE");
        code = taos_errno(res2);
399 400 401
        if (code != TSDB_CODE_SUCCESS) {
          tscError("SML:0x%" PRIx64 " apply schema action. reset query cache. error: %s", info->id, taos_errstr(res2));
        }
402 403
        taos_free_result(res2);
      }
404
      break;
S
shenglian zhou 已提交
405 406 407 408 409 410 411
    }
    case SCHEMA_ACTION_CHANGE_TAG_SIZE: {
      int n = sprintf(result, "alter stable %s modify tag ", action->alterSTable.sTableName);
      buildColumnDescription(action->alterSTable.field, result+n,
                             capacity-n, &outBytes);
      TAOS_RES* res = taos_query(taos, result); //TODO async doAsyncQuery
      code = taos_errno(res);
S
shenglian zhou 已提交
412 413 414
      if (code != TSDB_CODE_SUCCESS) {
        tscError("SML:0x%"PRIx64" apply schema action. error : %s", info->id, taos_errstr(res));
      }
415 416
      taos_free_result(res);

417
      if (code == TSDB_CODE_MND_INVALID_TAG_LENGTH || code == TSDB_CODE_TSC_INVALID_TAG_LENGTH) {
418 419
        TAOS_RES* res2 = taos_query(taos, "RESET QUERY CACHE");
        code = taos_errno(res2);
420 421 422
        if (code != TSDB_CODE_SUCCESS) {
          tscError("SML:0x%" PRIx64 " apply schema action. reset query cache. error: %s", info->id, taos_errstr(res2));
        }
423 424
        taos_free_result(res2);
      }
S
shenglian zhou 已提交
425 426 427 428 429
      break;
    }
    case SCHEMA_ACTION_CREATE_STABLE: {
      int n = sprintf(result, "create stable %s (", action->createSTable.sTableName);
      char* pos = result + n; int freeBytes = capacity - n;
S
shenglian zhou 已提交
430
      size_t numCols = taosArrayGetSize(action->createSTable.fields);
S
shenglian zhou 已提交
431 432 433 434 435 436 437
      for (int32_t i = 0; i < numCols; ++i) {
        SSchema* field = taosArrayGet(action->createSTable.fields, i);
        buildColumnDescription(field, pos, freeBytes, &outBytes);
        pos += outBytes; freeBytes -= outBytes;
        *pos = ','; ++pos; --freeBytes;
      }
      --pos; ++freeBytes;
438

S
shenglian zhou 已提交
439 440
      outBytes = snprintf(pos, freeBytes, ") tags (");
      pos += outBytes; freeBytes -= outBytes;
441

S
shenglian zhou 已提交
442
      size_t numTags = taosArrayGetSize(action->createSTable.tags);
S
shenglian zhou 已提交
443 444 445 446 447 448 449 450 451 452
      for (int32_t i = 0; i < numTags; ++i) {
        SSchema* field = taosArrayGet(action->createSTable.tags, i);
        buildColumnDescription(field, pos, freeBytes, &outBytes);
        pos += outBytes; freeBytes -= outBytes;
        *pos = ','; ++pos; --freeBytes;
      }
      pos--; ++freeBytes;
      outBytes = snprintf(pos, freeBytes, ")");
      TAOS_RES* res = taos_query(taos, result);
      code = taos_errno(res);
S
shenglian zhou 已提交
453 454 455
      if (code != TSDB_CODE_SUCCESS) {
        tscError("SML:0x%"PRIx64" apply schema action. error : %s", info->id, taos_errstr(res));
      }
456 457
      taos_free_result(res);

458 459 460
      if (code == TSDB_CODE_MND_TABLE_ALREADY_EXIST) {
        TAOS_RES* res2 = taos_query(taos, "RESET QUERY CACHE");
        code = taos_errno(res2);
461 462 463
        if (code != TSDB_CODE_SUCCESS) {
          tscError("SML:0x%" PRIx64 " apply schema action. reset query cache. error: %s", info->id, taos_errstr(res2));
        }
464 465
        taos_free_result(res2);
      }
S
shenglian zhou 已提交
466 467
      break;
    }
S
shenglian zhou 已提交
468

S
shenglian zhou 已提交
469 470 471
    default:
      break;
  }
S
Shenglian Zhou 已提交
472

S
shenglian zhou 已提交
473
  free(result);
S
Shenglian Zhou 已提交
474
  if (code != 0) {
S
shenglian zhou 已提交
475
    tscError("SML:0x%"PRIx64 " apply schema action failure. %s", info->id, tstrerror(code));
S
Shenglian Zhou 已提交
476
  }
S
shenglian zhou 已提交
477 478 479
  return code;
}

480
static int32_t destroySmlSTableSchema(SSmlSTableSchema* schema) {
S
shenglian zhou 已提交
481 482 483 484 485 486 487
  taosHashCleanup(schema->tagHash);
  taosHashCleanup(schema->fieldHash);
  taosArrayDestroy(schema->tags);
  taosArrayDestroy(schema->fields);
  return 0;
}

488
static int32_t fillDbSchema(STableMeta* tableMeta, char* tableName, SSmlSTableSchema* schema, SSmlLinesInfo* info) {
S
shenglian zhou 已提交
489 490 491 492 493 494 495 496 497 498 499 500
  schema->tags = taosArrayInit(8, sizeof(SSchema));
  schema->fields = taosArrayInit(64, sizeof(SSchema));
  schema->tagHash = taosHashInit(8, taosGetDefaultHashFunction(TSDB_DATA_TYPE_BINARY), true, false);
  schema->fieldHash = taosHashInit(64, taosGetDefaultHashFunction(TSDB_DATA_TYPE_BINARY), true, false);

  tstrncpy(schema->sTableName, tableName, strlen(tableName)+1);
  schema->precision = tableMeta->tableInfo.precision;
  for (int i=0; i<tableMeta->tableInfo.numOfColumns; ++i) {
    SSchema field;
    tstrncpy(field.name, tableMeta->schema[i].name, strlen(tableMeta->schema[i].name)+1);
    field.type = tableMeta->schema[i].type;
    field.bytes = tableMeta->schema[i].bytes;
501 502 503
    taosArrayPush(schema->fields, &field);
    size_t fieldIndex = taosArrayGetSize(schema->fields) - 1;
    taosHashPut(schema->fieldHash, field.name, strlen(field.name), &fieldIndex, sizeof(fieldIndex));
S
shenglian zhou 已提交
504 505 506 507 508 509 510 511
  }

  for (int i=0; i<tableMeta->tableInfo.numOfTags; ++i) {
    int j = i + tableMeta->tableInfo.numOfColumns;
    SSchema field;
    tstrncpy(field.name, tableMeta->schema[j].name, strlen(tableMeta->schema[j].name)+1);
    field.type = tableMeta->schema[j].type;
    field.bytes = tableMeta->schema[j].bytes;
512 513 514
    taosArrayPush(schema->tags, &field);
    size_t tagIndex = taosArrayGetSize(schema->tags) - 1;
    taosHashPut(schema->tagHash, field.name, strlen(field.name), &tagIndex, sizeof(tagIndex));
S
shenglian zhou 已提交
515
  }
516
  tscDebug("SML:0x%"PRIx64 " load table schema succeed. table name: %s, columns number: %d, tag number: %d, precision: %d",
S
shenglian zhou 已提交
517
           info->id, tableName, tableMeta->tableInfo.numOfColumns, tableMeta->tableInfo.numOfTags, schema->precision);
518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 556 557
  return TSDB_CODE_SUCCESS;
}

static int32_t retrieveTableMeta(TAOS* taos, char* tableName, STableMeta** pTableMeta, SSmlLinesInfo* info) {
  int32_t code = 0;
  int32_t retries = 0;
  STableMeta* tableMeta = NULL;
  while (retries++ < TSDB_MAX_REPLICA && tableMeta == NULL) {
    STscObj* pObj = (STscObj*)taos;
    if (pObj == NULL || pObj->signature != pObj) {
      terrno = TSDB_CODE_TSC_DISCONNECTED;
      return TSDB_CODE_TSC_DISCONNECTED;
    }

    tscDebug("SML:0x%" PRIx64 " retrieve table meta. super table name: %s", info->id, tableName);

    char tableNameLowerCase[TSDB_TABLE_NAME_LEN];
    strtolower(tableNameLowerCase, tableName);

    char sql[256];
    snprintf(sql, 256, "describe %s", tableNameLowerCase);
    TAOS_RES* res = taos_query(taos, sql);
    code = taos_errno(res);
    if (code != 0) {
      tscError("SML:0x%" PRIx64 " describe table failure. %s", info->id, taos_errstr(res));
      taos_free_result(res);
      return code;
    }
    taos_free_result(res);

    SSqlObj* pSql = calloc(1, sizeof(SSqlObj));
    if (pSql == NULL) {
      tscError("SML:0x%" PRIx64 " failed to allocate memory, reason:%s", info->id, strerror(errno));
      code = TSDB_CODE_TSC_OUT_OF_MEMORY;
      return code;
    }
    pSql->pTscObj = taos;
    pSql->signature = pSql;
    pSql->fp = NULL;

558
    registerSqlObj(pSql);
559 560 561 562 563 564
    SStrToken tableToken = {.z = tableNameLowerCase, .n = (uint32_t)strlen(tableNameLowerCase), .type = TK_ID};
    tGetToken(tableNameLowerCase, &tableToken.type);
    // Check if the table name available or not
    if (tscValidateName(&tableToken) != TSDB_CODE_SUCCESS) {
      code = TSDB_CODE_TSC_INVALID_TABLE_ID_LENGTH;
      sprintf(pSql->cmd.payload, "table name is invalid");
565
      tscFreeRegisteredSqlObj(pSql);
566 567 568 569 570
      return code;
    }

    SName sname = {0};
    if ((code = tscSetTableFullName(&sname, &tableToken, pSql)) != TSDB_CODE_SUCCESS) {
571
      tscFreeRegisteredSqlObj(pSql);
572 573 574 575 576 577 578
      return code;
    }
    char fullTableName[TSDB_TABLE_FNAME_LEN] = {0};
    memset(fullTableName, 0, tListLen(fullTableName));
    tNameExtractFullName(&sname, fullTableName);

    size_t size = 0;
579
    taosHashGetCloneExt(UTIL_GET_TABLEMETA(pSql), fullTableName, strlen(fullTableName), NULL, (void**)&tableMeta, &size);
580
    tscFreeRegisteredSqlObj(pSql);
581 582 583 584 585 586 587 588 589 590 591 592 593 594 595 596 597 598 599 600 601 602
  }

  if (tableMeta != NULL) {
    *pTableMeta = tableMeta;
    return TSDB_CODE_SUCCESS;
  } else {
    tscError("SML:0x%" PRIx64 " failed to retrieve table meta. super table name: %s", info->id, tableName);
    return TSDB_CODE_TSC_NO_META_CACHED;
  }
}

static int32_t loadTableSchemaFromDB(TAOS* taos, char* tableName, SSmlSTableSchema* schema, SSmlLinesInfo* info) {
  int32_t code = 0;
  STableMeta* tableMeta = NULL;
  code = retrieveTableMeta(taos, tableName, &tableMeta, info);
  if (code == TSDB_CODE_SUCCESS) {
    assert(tableMeta != NULL);
    fillDbSchema(tableMeta, tableName, schema, info);
    free(tableMeta);
    tableMeta = NULL;
  }

S
shenglian zhou 已提交
603 604 605
  return code;
}

S
shenglian zhou 已提交
606
static int32_t modifyDBSchemas(TAOS* taos, SArray* stableSchemas, SSmlLinesInfo* info) {
S
shenglian zhou 已提交
607 608 609 610
  int32_t code = 0;
  size_t numStable = taosArrayGetSize(stableSchemas);
  for (int i = 0; i < numStable; ++i) {
    SSmlSTableSchema* pointSchema = taosArrayGet(stableSchemas, i);
S
shenglian zhou 已提交
611 612
    SSmlSTableSchema  dbSchema;
    memset(&dbSchema, 0, sizeof(SSmlSTableSchema));
S
shenglian zhou 已提交
613

614
    code = loadTableSchemaFromDB(taos, pointSchema->sTableName, &dbSchema, info);
S
shenglian zhou 已提交
615 616 617 618 619 620 621
    if (code == TSDB_CODE_MND_INVALID_TABLE_NAME) {
      SSchemaAction schemaAction = {0};
      schemaAction.action = SCHEMA_ACTION_CREATE_STABLE;
      memset(&schemaAction.createSTable, 0, sizeof(SCreateSTableActionInfo));
      memcpy(schemaAction.createSTable.sTableName, pointSchema->sTableName, TSDB_TABLE_NAME_LEN);
      schemaAction.createSTable.tags = pointSchema->tags;
      schemaAction.createSTable.fields = pointSchema->fields;
S
shenglian zhou 已提交
622
      applySchemaAction(taos, &schemaAction, info);
623
      code = loadTableSchemaFromDB(taos, pointSchema->sTableName, &dbSchema, info);
S
Shenglian Zhou 已提交
624
      if (code != 0) {
S
shenglian zhou 已提交
625
        tscError("SML:0x%"PRIx64" reconcile point schema failed. can not create %s", info->id, pointSchema->sTableName);
626
        return code;
S
Shenglian Zhou 已提交
627 628 629 630
      } else {
        pointSchema->precision = dbSchema.precision;
        destroySmlSTableSchema(&dbSchema);
      }
S
shenglian zhou 已提交
631 632 633 634 635 636 637 638 639 640 641
    } else if (code == TSDB_CODE_SUCCESS) {
      size_t pointTagSize = taosArrayGetSize(pointSchema->tags);
      size_t pointFieldSize = taosArrayGetSize(pointSchema->fields);

      SHashObj* dbTagHash = dbSchema.tagHash;
      SHashObj* dbFieldHash = dbSchema.fieldHash;

      for (int j = 0; j < pointTagSize; ++j) {
        SSchema* pointTag = taosArrayGet(pointSchema->tags, j);
        SSchemaAction schemaAction = {0};
        bool actionNeeded = false;
S
shenglian zhou 已提交
642 643
        generateSchemaAction(pointTag, dbTagHash, dbSchema.tags, true, pointSchema->sTableName,
                             &schemaAction, &actionNeeded, info);
S
shenglian zhou 已提交
644
        if (actionNeeded) {
S
shenglian zhou 已提交
645
          code = applySchemaAction(taos, &schemaAction, info);
646 647 648 649
          if (code != 0) {
            destroySmlSTableSchema(&dbSchema);
            return code;
          }
S
shenglian zhou 已提交
650 651 652 653 654 655 656 657 658 659 660
        }
      }

      SSchema* pointColTs = taosArrayGet(pointSchema->fields, 0);
      SSchema* dbColTs = taosArrayGet(dbSchema.fields, 0);
      memcpy(pointColTs->name, dbColTs->name, TSDB_COL_NAME_LEN);

      for (int j = 1; j < pointFieldSize; ++j) {
        SSchema* pointCol = taosArrayGet(pointSchema->fields, j);
        SSchemaAction schemaAction = {0};
        bool actionNeeded = false;
S
shenglian zhou 已提交
661 662
        generateSchemaAction(pointCol, dbFieldHash, dbSchema.fields,false, pointSchema->sTableName,
                             &schemaAction, &actionNeeded, info);
S
shenglian zhou 已提交
663
        if (actionNeeded) {
S
shenglian zhou 已提交
664
          code = applySchemaAction(taos, &schemaAction, info);
665 666 667 668
          if (code != 0) {
            destroySmlSTableSchema(&dbSchema);
            return code;
          }
S
shenglian zhou 已提交
669 670 671 672 673
        }
      }

      pointSchema->precision = dbSchema.precision;

674
      destroySmlSTableSchema(&dbSchema);
S
shenglian zhou 已提交
675
    } else {
S
shenglian zhou 已提交
676
      tscError("SML:0x%"PRIx64" load table meta error: %s", info->id, tstrerror(code));
S
shenglian zhou 已提交
677 678 679 680 681 682
      return code;
    }
  }
  return 0;
}

S
shenglian zhou 已提交
683 684
static int32_t creatChildTableIfNotExists(TAOS* taos, const char* cTableName, const char* sTableName,
                                          SArray* tagsSchema, SArray* tagsBind, SSmlLinesInfo* info) {
S
shenglian zhou 已提交
685
  size_t numTags = taosArrayGetSize(tagsSchema);
686
  char* sql = malloc(tsMaxSQLStringLen+1);
687 688 689 690
  if (sql == NULL) {
    tscError("malloc sql memory error");
    return TSDB_CODE_TSC_OUT_OF_MEMORY;
  }
691
  int freeBytes = tsMaxSQLStringLen + 1;
S
shenglian zhou 已提交
692
  sprintf(sql, "create table if not exists %s using %s", cTableName, sTableName);
693

S
shenglian zhou 已提交
694 695 696 697 698 699
  snprintf(sql+strlen(sql), freeBytes-strlen(sql), "(");
  for (int i = 0; i < numTags; ++i) {
    SSchema* tagSchema = taosArrayGet(tagsSchema, i);
    snprintf(sql+strlen(sql), freeBytes-strlen(sql), "%s,", tagSchema->name);
  }
  snprintf(sql + strlen(sql) - 1, freeBytes-strlen(sql)+1, ")");
700

S
shenglian zhou 已提交
701
  snprintf(sql + strlen(sql), freeBytes-strlen(sql), " tags (");
702

S
shenglian zhou 已提交
703
  for (int i = 0; i < numTags; ++i) {
S
shenglian zhou 已提交
704
    snprintf(sql+strlen(sql), freeBytes-strlen(sql), "?,");
S
shenglian zhou 已提交
705
  }
S
shenglian zhou 已提交
706
  snprintf(sql + strlen(sql) - 1, freeBytes-strlen(sql)+1, ")");
707
  sql[strlen(sql)] = '\0';
S
shenglian zhou 已提交
708

S
shenglian zhou 已提交
709
  tscDebug("SML:0x%"PRIx64" create table : %s", info->id, sql);
S
Shenglian Zhou 已提交
710

S
shenglian zhou 已提交
711
  TAOS_STMT* stmt = taos_stmt_init(taos);
712 713 714 715
  if (stmt == NULL) {
    free(sql);
    return TSDB_CODE_TSC_OUT_OF_MEMORY;
  }
S
shenglian zhou 已提交
716
  int32_t code;
S
shenglian zhou 已提交
717
  code = taos_stmt_prepare(stmt, sql, (unsigned long)strlen(sql));
718 719
  free(sql);

S
shenglian zhou 已提交
720
  if (code != 0) {
721
    tscError("SML:0x%"PRIx64" taos_stmt_prepare returns %d:%s", info->id, code, tstrerror(code));
722
    taos_stmt_close(stmt);
S
shenglian zhou 已提交
723 724 725 726 727
    return code;
  }

  code = taos_stmt_bind_param(stmt, TARRAY_GET_START(tagsBind));
  if (code != 0) {
728
    tscError("SML:0x%"PRIx64" taos_stmt_bind_param returns %d:%s", info->id, code, tstrerror(code));
729
    taos_stmt_close(stmt);
S
shenglian zhou 已提交
730 731 732 733 734
    return code;
  }

  code = taos_stmt_execute(stmt);
  if (code != 0) {
735
    tscError("SML:0x%"PRIx64" taos_stmt_execute returns %d:%s", info->id, code, tstrerror(code));
736
    taos_stmt_close(stmt);
S
shenglian zhou 已提交
737 738
    return code;
  }
S
shenglian zhou 已提交
739

740 741
  code = taos_stmt_close(stmt);
  if (code != 0) {
742
    tscError("SML:0x%"PRIx64" taos_stmt_close return %d:%s", info->id, code, tstrerror(code));
743 744 745
    return code;
  }
  return code;
S
shenglian zhou 已提交
746 747
}

S
shenglian zhou 已提交
748
static int32_t insertChildTableBatch(TAOS* taos,  char* cTableName, SArray* colsSchema, SArray* rowsBind, SSmlLinesInfo* info) {
S
shenglian zhou 已提交
749
  size_t numCols = taosArrayGetSize(colsSchema);
750
  char* sql = malloc(tsMaxSQLStringLen+1);
751 752 753 754 755
  if (sql == NULL) {
    tscError("malloc sql memory error");
    return TSDB_CODE_TSC_OUT_OF_MEMORY;
  }

756
  int32_t freeBytes = tsMaxSQLStringLen + 1 ;
S
shenglian zhou 已提交
757
  sprintf(sql, "insert into ? (");
758

S
shenglian zhou 已提交
759 760
  for (int i = 0; i < numCols; ++i) {
    SSchema* colSchema = taosArrayGet(colsSchema, i);
S
shenglian zhou 已提交
761
    snprintf(sql+strlen(sql), freeBytes-strlen(sql), "%s,", colSchema->name);
S
shenglian zhou 已提交
762
  }
S
shenglian zhou 已提交
763
  snprintf(sql + strlen(sql)-1, freeBytes-strlen(sql)+1, ") values (");
764

S
shenglian zhou 已提交
765
  for (int i = 0; i < numCols; ++i) {
S
shenglian zhou 已提交
766
    snprintf(sql+strlen(sql), freeBytes-strlen(sql), "?,");
767
  }
S
shenglian zhou 已提交
768
  snprintf(sql + strlen(sql)-1, freeBytes-strlen(sql)+1, ")");
769
  sql[strlen(sql)] = '\0';
770

S
shenglian zhou 已提交
771
  tscDebug("SML:0x%"PRIx64" insert rows into child table %s. num of rows: %zu", info->id, cTableName, taosArrayGetSize(rowsBind));
W
wpan 已提交
772

773
  int32_t code = 0;
S
shenglian zhou 已提交
774

W
wpan 已提交
775
  TAOS_STMT* stmt = taos_stmt_init(taos);
776
  if (stmt == NULL) {
777
    tfree(sql);
778 779
    return TSDB_CODE_TSC_OUT_OF_MEMORY;
  }
780

S
shenglian zhou 已提交
781
  code = taos_stmt_prepare(stmt, sql, (unsigned long)strlen(sql));
782
  tfree(sql);
783

W
wpan 已提交
784
  if (code != 0) {
785
    tscError("SML:0x%"PRIx64" taos_stmt_prepare return %d:%s", info->id, code, tstrerror(code));
786
    taos_stmt_close(stmt);
W
wpan 已提交
787 788 789
    return code;
  }

790
  bool tryAgain = false;
791
  int32_t try = 0;
W
wpan 已提交
792
  do {
793
    code = taos_stmt_set_tbname(stmt, cTableName);
794
    if (code != 0) {
795
      tscError("SML:0x%"PRIx64" taos_stmt_set_tbname return %d:%s", info->id, code, tstrerror(code));
796
      taos_stmt_close(stmt);
797 798
      return code;
    }
799

800 801 802 803 804
    size_t rows = taosArrayGetSize(rowsBind);
    for (int32_t i = 0; i < rows; ++i) {
      TAOS_BIND* colsBinds = taosArrayGetP(rowsBind, i);
      code = taos_stmt_bind_param(stmt, colsBinds);
      if (code != 0) {
805
        tscError("SML:0x%"PRIx64" taos_stmt_bind_param return %d:%s", info->id, code, tstrerror(code));
806
        taos_stmt_close(stmt);
807 808 809 810
        return code;
      }
      code = taos_stmt_add_batch(stmt);
      if (code != 0) {
811
        tscError("SML:0x%"PRIx64" taos_stmt_add_batch return %d:%s", info->id, code, tstrerror(code));
812
        taos_stmt_close(stmt);
813 814 815 816 817 818
        return code;
      }
    }

    code = taos_stmt_execute(stmt);
    if (code != 0) {
819
      tscError("SML:0x%"PRIx64" taos_stmt_execute return %d:%s, try:%d", info->id, code, tstrerror(code), try);
820
    }
821 822 823 824

    tryAgain = false;
    if ((code == TSDB_CODE_TDB_INVALID_TABLE_ID
        || code == TSDB_CODE_VND_INVALID_VGROUP_ID
825 826 827
        || code == TSDB_CODE_TDB_TABLE_RECONFIGURE
        || code == TSDB_CODE_APP_NOT_READY
        || code == TSDB_CODE_RPC_NETWORK_UNAVAIL) && try++ < TSDB_MAX_REPLICA) {
828 829 830
      tryAgain = true;
    }

831
    if (code == TSDB_CODE_TDB_INVALID_TABLE_ID || code == TSDB_CODE_VND_INVALID_VGROUP_ID) {
832 833 834 835 836 837
      TAOS_RES* res2 = taos_query(taos, "RESET QUERY CACHE");
      int32_t   code2 = taos_errno(res2);
      if (code2 != TSDB_CODE_SUCCESS) {
        tscError("SML:0x%" PRIx64 " insert child table. reset query cache. error: %s", info->id, taos_errstr(res2));
      }
      taos_free_result(res2);
838 839
      if (tryAgain) {
        taosMsleep(50 * (2 << try));
840
      }
841
    }
842
    if (code == TSDB_CODE_APP_NOT_READY || code == TSDB_CODE_RPC_NETWORK_UNAVAIL) {
843
      if (tryAgain) {
844
        taosMsleep( 50 * (2 << try));
845 846 847
      }
    }
  } while (tryAgain);
S
shenglian zhou 已提交
848

W
wpan 已提交
849

850
  taos_stmt_close(stmt);
S
shenglian zhou 已提交
851
  return code;
S
shenglian zhou 已提交
852 853
}

854
static int32_t arrangePointsByChildTableName(TAOS_SML_DATA_POINT* points, int numPoints,
S
shenglian zhou 已提交
855
                                             SHashObj* cname2points, SArray* stableSchemas, SSmlLinesInfo* info) {
S
shenglian zhou 已提交
856 857
  for (int32_t i = 0; i < numPoints; ++i) {
    TAOS_SML_DATA_POINT * point = points + i;
858 859 860 861
    uintptr_t valPointer = (uintptr_t)point;
    size_t* pSchemaIndex = taosHashGet(info->smlDataToSchema, &valPointer, sizeof(uintptr_t));
    assert(pSchemaIndex != NULL);
    SSmlSTableSchema* stableSchema = taosArrayGet(stableSchemas, *pSchemaIndex);
862

S
shenglian zhou 已提交
863 864 865 866
    for (int j = 0; j < point->tagNum; ++j) {
      TAOS_SML_KV* kv =  point->tags + j;
      if (kv->type == TSDB_DATA_TYPE_TIMESTAMP) {
        int64_t ts = *(int64_t*)(kv->value);
867
        ts = convertTimePrecision(ts, TSDB_TIME_PRECISION_NANO, stableSchema->precision);
S
shenglian zhou 已提交
868 869 870 871 872 873 874 875
        *(int64_t*)(kv->value) = ts;
      }
    }

    for (int j = 0; j < point->fieldNum; ++j) {
      TAOS_SML_KV* kv =  point->fields + j;
      if (kv->type == TSDB_DATA_TYPE_TIMESTAMP) {
        int64_t ts = *(int64_t*)(kv->value);
876
        ts = convertTimePrecision(ts, TSDB_TIME_PRECISION_NANO, stableSchema->precision);
S
shenglian zhou 已提交
877 878 879 880
        *(int64_t*)(kv->value) = ts;
      }
    }

S
shenglian zhou 已提交
881 882 883 884 885 886 887 888
    SArray* cTablePoints = NULL;
    SArray** pCTablePoints = taosHashGet(cname2points, point->childTableName, strlen(point->childTableName));
    if (pCTablePoints) {
      cTablePoints = *pCTablePoints;
    } else {
      cTablePoints = taosArrayInit(64, sizeof(point));
      taosHashPut(cname2points, point->childTableName, strlen(point->childTableName), &cTablePoints, POINTER_BYTES);
    }
889
    taosArrayPush(cTablePoints, &point);
890 891
  }

S
shenglian zhou 已提交
892 893 894
  return 0;
}

895
static int32_t applyChildTableTags(TAOS* taos, char* cTableName, char* sTableName,
S
shenglian zhou 已提交
896
                                   SSmlSTableSchema* sTableSchema, SArray* cTablePoints, SSmlLinesInfo* info) {
897 898 899 900 901 902 903 904
  size_t numTags = taosArrayGetSize(sTableSchema->tags);
  size_t rows = taosArrayGetSize(cTablePoints);

  TAOS_SML_KV* tagKVs[TSDB_MAX_TAGS] = {0};
  for (int i= 0; i < rows; ++i) {
    TAOS_SML_DATA_POINT * pDataPoint = taosArrayGetP(cTablePoints, i);
    for (int j = 0; j < pDataPoint->tagNum; ++j) {
      TAOS_SML_KV* kv = pDataPoint->tags + j;
905 906 907 908
      uintptr_t valPointer = (uintptr_t)kv;
      size_t* pFieldSchemaIdx = taosHashGet(info->smlDataToSchema, &valPointer, sizeof(uintptr_t));
      assert(pFieldSchemaIdx != NULL);
      tagKVs[*pFieldSchemaIdx] = kv;
909 910 911 912 913 914 915 916 917 918 919 920 921
    }
  }
  
  SArray* tagBinds = taosArrayInit(numTags, sizeof(TAOS_BIND));
  taosArraySetSize(tagBinds, numTags);
  int isNullColBind = TSDB_TRUE;
  for (int j = 0; j < numTags; ++j) {
    TAOS_BIND* bind = taosArrayGet(tagBinds, j);
    bind->is_null = &isNullColBind;
  }
  for (int j = 0; j < numTags; ++j) {
    if (tagKVs[j] == NULL) continue;
    TAOS_SML_KV* kv =  tagKVs[j];
922 923 924 925
    uintptr_t valPointer = (uintptr_t)kv;
    size_t* pFieldSchemaIdx = taosHashGet(info->smlDataToSchema, &valPointer, sizeof(uintptr_t));
    assert(pFieldSchemaIdx != NULL);
    TAOS_BIND* bind = taosArrayGet(tagBinds, *pFieldSchemaIdx);
926 927 928 929 930 931 932
    bind->buffer_type = kv->type;
    bind->length = malloc(sizeof(uintptr_t*));
    *bind->length = kv->length;
    bind->buffer = kv->value;
    bind->is_null = NULL;
  }

933
  int32_t code = creatChildTableIfNotExists(taos, cTableName, sTableName, sTableSchema->tags, tagBinds, info);
934 935 936 937 938 939 940 941 942

  for (int i = 0; i < taosArrayGetSize(tagBinds); ++i) {
    TAOS_BIND* bind = taosArrayGet(tagBinds, i);
    free(bind->length);
  }
  taosArrayDestroy(tagBinds);
  return code;
}

S
shenglian zhou 已提交
943 944
static int32_t applyChildTableFields(TAOS* taos, SSmlSTableSchema* sTableSchema, char* cTableName,
                                     SArray* cTablePoints, SSmlLinesInfo* info) {
945 946
  int32_t code = TSDB_CODE_SUCCESS;

947 948 949
  size_t numCols = taosArrayGetSize(sTableSchema->fields);
  size_t rows = taosArrayGetSize(cTablePoints);
  SArray* rowsBind = taosArrayInit(rows, POINTER_BYTES);
S
shenglian zhou 已提交
950

951 952
  for (int i = 0; i < rows; ++i) {
    TAOS_SML_DATA_POINT* point = taosArrayGetP(cTablePoints, i);
953

954 955
    TAOS_BIND* colBinds = calloc(numCols, sizeof(TAOS_BIND));
    if (colBinds == NULL) {
S
shenglian zhou 已提交
956 957
      tscError("SML:0x%"PRIx64" taos_sml_insert insert points, failed to allocated memory for TAOS_BIND, "
               "num of rows: %zu, num of cols: %zu", info->id, rows, numCols);
958 959
      return TSDB_CODE_TSC_OUT_OF_MEMORY;
    }
960

961 962 963
    int isNullColBind = TSDB_TRUE;
    for (int j = 0; j < numCols; ++j) {
      TAOS_BIND* bind = colBinds + j;
964 965
      bind->is_null = &isNullColBind;
    }
966 967
    for (int j = 0; j < point->fieldNum; ++j) {
      TAOS_SML_KV* kv = point->fields + j;
968 969 970 971
      uintptr_t valPointer = (uintptr_t)kv;
      size_t* pFieldSchemaIdx = taosHashGet(info->smlDataToSchema, &valPointer, sizeof(uintptr_t));
      assert(pFieldSchemaIdx != NULL);
      TAOS_BIND* bind = colBinds + *pFieldSchemaIdx;
972
      bind->buffer_type = kv->type;
973 974
      bind->length = malloc(sizeof(uintptr_t*));
      *bind->length = kv->length;
975 976 977
      bind->buffer = kv->value;
      bind->is_null = NULL;
    }
978 979
    taosArrayPush(rowsBind, &colBinds);
  }
S
shenglian zhou 已提交
980

981 982 983 984
  code = insertChildTableBatch(taos, cTableName, sTableSchema->fields, rowsBind, info);
  if (code != 0) {
    tscError("SML:0x%"PRIx64" insert into child table %s failed. error %s", info->id, cTableName, tstrerror(code));
  }
S
shenglian zhou 已提交
985

986 987 988 989 990
  for (int i = 0; i < rows; ++i) {
    TAOS_BIND* colBinds = taosArrayGetP(rowsBind, i);
    for (int j = 0; j < numCols; ++j) {
      TAOS_BIND* bind = colBinds + j;
      free(bind->length);
S
shenglian zhou 已提交
991
    }
992 993 994 995 996
    free(colBinds);
  }
  taosArrayDestroy(rowsBind);
  return code;
}
S
shenglian zhou 已提交
997

S
shenglian zhou 已提交
998
static int32_t applyDataPoints(TAOS* taos, TAOS_SML_DATA_POINT* points, int32_t numPoints, SArray* stableSchemas, SSmlLinesInfo* info) {
999
  int32_t code = TSDB_CODE_SUCCESS;
1000

1001
  SHashObj* cname2points = taosHashInit(128, taosGetDefaultHashFunction(TSDB_DATA_TYPE_BINARY), true, false);
S
shenglian zhou 已提交
1002
  arrangePointsByChildTableName(points, numPoints, cname2points, stableSchemas, info);
1003 1004 1005 1006 1007 1008

  SArray** pCTablePoints = taosHashIterate(cname2points, NULL);
  while (pCTablePoints) {
    SArray* cTablePoints = *pCTablePoints;

    TAOS_SML_DATA_POINT* point = taosArrayGetP(cTablePoints, 0);
1009 1010 1011 1012
    uintptr_t valPointer = (uintptr_t)point;
    size_t* pSchemaIndex = taosHashGet(info->smlDataToSchema, &valPointer, sizeof(uintptr_t));
    assert(pSchemaIndex != NULL);
    SSmlSTableSchema*    sTableSchema = taosArrayGet(stableSchemas, *pSchemaIndex);
S
shenglian zhou 已提交
1013 1014 1015

    tscDebug("SML:0x%"PRIx64" apply child table tags. child table: %s", info->id, point->childTableName);
    code = applyChildTableTags(taos, point->childTableName, point->stableName, sTableSchema, cTablePoints, info);
1016 1017 1018
    if (code != 0) {
      tscError("apply child table tags failed. child table %s, error %s", point->childTableName, tstrerror(code));
      goto cleanup;
S
shenglian zhou 已提交
1019
    }
S
shenglian zhou 已提交
1020 1021 1022

    tscDebug("SML:0x%"PRIx64" apply child table points. child table: %s", info->id, point->childTableName);
    code = applyChildTableFields(taos, sTableSchema, point->childTableName, cTablePoints, info);
1023
    if (code != 0) {
1024
      tscError("SML:0x%"PRIx64" Apply child table fields failed. child table %s, error %s", info->id, point->childTableName, tstrerror(code));
1025
      goto cleanup;
1026
    }
1027

S
shenglian zhou 已提交
1028
    tscDebug("SML:0x%"PRIx64" successfully applied data points of child table %s", info->id, point->childTableName);
S
shenglian zhou 已提交
1029

S
shenglian zhou 已提交
1030
    pCTablePoints = taosHashIterate(cname2points, pCTablePoints);
1031
  }
S
shenglian zhou 已提交
1032

1033 1034 1035 1036 1037
cleanup:
  pCTablePoints = taosHashIterate(cname2points, NULL);
  while (pCTablePoints) {
    SArray* pPoints = *pCTablePoints;
    taosArrayDestroy(pPoints);
S
shenglian zhou 已提交
1038
    pCTablePoints = taosHashIterate(cname2points, pCTablePoints);
1039
  }
S
shenglian zhou 已提交
1040
  taosHashCleanup(cname2points);
1041
  return code;
S
shenglian zhou 已提交
1042
}
1043

1044
int tscSmlInsert(TAOS* taos, TAOS_SML_DATA_POINT* points, int numPoint, SSmlLinesInfo* info) {
S
shenglian zhou 已提交
1045
  tscDebug("SML:0x%"PRIx64" taos_sml_insert. number of points: %d", info->id, numPoint);
S
Shenglian Zhou 已提交
1046

S
shenglian zhou 已提交
1047
  int32_t code = TSDB_CODE_SUCCESS;
1048
  info->smlDataToSchema = taosHashInit(32, taosGetDefaultHashFunction(TSDB_DATA_TYPE_UBIGINT), true, false);
S
shenglian zhou 已提交
1049

S
shenglian zhou 已提交
1050
  tscDebug("SML:0x%"PRIx64" build data point schemas", info->id);
S
shenglian zhou 已提交
1051
  SArray* stableSchemas = taosArrayInit(32, sizeof(SSmlSTableSchema)); // SArray<STableColumnsSchema>
S
shenglian zhou 已提交
1052
  code = buildDataPointSchemas(points, numPoint, stableSchemas, info);
S
shenglian zhou 已提交
1053
  if (code != 0) {
S
shenglian zhou 已提交
1054
    tscError("SML:0x%"PRIx64" error building data point schemas : %s", info->id, tstrerror(code));
S
shenglian zhou 已提交
1055 1056 1057
    goto clean_up;
  }

S
shenglian zhou 已提交
1058 1059
  tscDebug("SML:0x%"PRIx64" modify db schemas", info->id);
  code = modifyDBSchemas(taos, stableSchemas, info);
S
shenglian zhou 已提交
1060
  if (code != 0) {
S
shenglian zhou 已提交
1061
    tscError("SML:0x%"PRIx64" error change db schema : %s", info->id, tstrerror(code));
S
shenglian zhou 已提交
1062 1063
    goto clean_up;
  }
S
shenglian zhou 已提交
1064

S
shenglian zhou 已提交
1065 1066
  tscDebug("SML:0x%"PRIx64" apply data points", info->id);
  code = applyDataPoints(taos, points, numPoint, stableSchemas, info);
S
shenglian zhou 已提交
1067
  if (code != 0) {
S
shenglian zhou 已提交
1068
    tscError("SML:0x%"PRIx64" error apply data points : %s", info->id, tstrerror(code));
S
shenglian zhou 已提交
1069 1070 1071
  }

clean_up:
S
shenglian zhou 已提交
1072 1073 1074 1075
  for (int i = 0; i < taosArrayGetSize(stableSchemas); ++i) {
    SSmlSTableSchema* schema = taosArrayGet(stableSchemas, i);
    taosArrayDestroy(schema->fields);
    taosArrayDestroy(schema->tags);
1076
  }
S
shenglian zhou 已提交
1077
  taosArrayDestroy(stableSchemas);
1078 1079 1080 1081 1082 1083 1084 1085 1086
  taosHashCleanup(info->smlDataToSchema);
  return code;
}

int taos_sml_insert(TAOS* taos, TAOS_SML_DATA_POINT* points, int numPoint) {
  SSmlLinesInfo* info = calloc(1, sizeof(SSmlLinesInfo));
  info->id = genLinesSmlId();
  int code = tscSmlInsert(taos, points, numPoint, info);
  free(info);
1087 1088
  return code;
}
S
shenglian zhou 已提交
1089

1090 1091
//=========================================================================

1092 1093 1094 1095 1096
/*        Field                          Escape charaters
    1: measurement                        Comma,Space
    2: tag_key, tag_value, field_key  Comma,Equal Sign,Space
    3: field_value                    Double quote,Backslash
*/
1097
static void escapeSpecialCharacter(uint8_t field, const char **pos) {
1098 1099 1100
  const char *cur = *pos;
  if (*cur != '\\') {
    return;
1101
  }
1102 1103 1104 1105 1106 1107
  switch (field) {
    case 1:
      switch (*(cur + 1)) {
        case ',':
        case ' ':
          cur++;
1108
          break;
1109 1110 1111 1112 1113 1114 1115 1116 1117 1118 1119 1120 1121 1122 1123 1124 1125 1126 1127 1128 1129 1130 1131 1132 1133 1134 1135
        default:
          break;
      }
      break;
    case 2:
      switch (*(cur + 1)) {
        case ',':
        case ' ':
        case '=':
          cur++;
          break;
        default:
          break;
      }
      break;
    case 3:
      switch (*(cur + 1)) {
        case '"':
        case '\\':
          cur++;
          break;
        default:
          break;
      }
      break;
    default:
      break;
1136
  }
1137
  *pos = cur;
1138
}
1139

1140
bool isValidInteger(char *str) {
1141 1142
  char *c = str;
  if (*c != '+' && *c != '-' && !isdigit(*c)) {
1143 1144
    return false;
  }
1145 1146 1147 1148 1149 1150
  c++;
  while (*c != '\0') {
    if (!isdigit(*c)) {
      return false;
    }
    c++;
1151
  }
1152
  return true;
1153
}
1154

1155
bool isValidFloat(char *str) {
1156 1157 1158 1159 1160 1161 1162
  char *c = str;
  uint8_t has_dot, has_exp, has_sign;
  has_dot = 0;
  has_exp = 0;
  has_sign = 0;

  if (*c != '+' && *c != '-' && *c != '.' && !isdigit(*c)) {
1163 1164
    return false;
  }
1165 1166
  if (*c == '.' && isdigit(*(c + 1))) {
    has_dot = 1;
1167
  }
1168 1169 1170 1171 1172 1173 1174 1175 1176 1177 1178 1179 1180 1181 1182 1183 1184 1185 1186 1187 1188 1189 1190 1191 1192 1193 1194 1195 1196 1197 1198 1199 1200 1201 1202 1203 1204 1205 1206 1207 1208
  c++;
  while (*c != '\0') {
    if (!isdigit(*c)) {
      switch (*c) {
        case '.': {
          if (!has_dot && !has_exp && isdigit(*(c + 1))) {
            has_dot = 1;
          } else {
            return false;
          }
          break;
        }
        case 'e':
        case 'E': {
          if (!has_exp && isdigit(*(c - 1)) &&
              (isdigit(*(c + 1)) ||
               *(c + 1) == '+' ||
               *(c + 1) == '-')) {
            has_exp = 1;
          } else {
            return false;
          }
          break;
        }
        case '+':
        case '-': {
          if (!has_sign && has_exp && isdigit(*(c + 1))) {
            has_sign = 1;
          } else {
            return false;
          }
          break;
        }
        default: {
          return false;
        }
      }
    }
    c++;
  } //while
  return true;
1209
}
1210

1211 1212 1213 1214 1215 1216 1217 1218 1219 1220 1221 1222 1223 1224 1225 1226
static bool isInteger(char *pVal, uint16_t len, bool *has_sign) {
  if (len <= 1) {
    return false;
  }
  if (pVal[len - 1] == 'i') {
    *has_sign = true;
    return true;
  }
  if (pVal[len - 1] == 'u') {
    *has_sign = false;
    return true;
  }

  return false;
}

1227
static bool isTinyInt(char *pVal, uint16_t len) {
1228 1229 1230
  if (len <= 2) {
    return false;
  }
1231
  if (!strcasecmp(&pVal[len - 2], "i8")) {
1232
    //printf("Type is int8(%s)\n", pVal);
1233 1234 1235 1236
    return true;
  }
  return false;
}
1237

1238
static bool isTinyUint(char *pVal, uint16_t len) {
1239 1240 1241 1242 1243
  if (len <= 2) {
    return false;
  }
  if (pVal[0] == '-') {
    return false;
1244
  }
1245
  if (!strcasecmp(&pVal[len - 2], "u8")) {
1246
    //printf("Type is uint8(%s)\n", pVal);
1247 1248 1249
    return true;
  }
  return false;
1250 1251
}

1252
static bool isSmallInt(char *pVal, uint16_t len) {
1253 1254 1255
  if (len <= 3) {
    return false;
  }
1256
  if (!strcasecmp(&pVal[len - 3], "i16")) {
1257
    //printf("Type is int16(%s)\n", pVal);
1258
    return true;
1259
  }
1260
  return false;
1261 1262
}

1263
static bool isSmallUint(char *pVal, uint16_t len) {
1264 1265
  if (len <= 3) {
    return false;
1266
  }
1267 1268 1269
  if (pVal[0] == '-') {
    return false;
  }
1270
  if (strcasecmp(&pVal[len - 3], "u16") == 0) {
1271
    //printf("Type is uint16(%s)\n", pVal);
1272 1273 1274
    return true;
  }
  return false;
1275 1276
}

1277
static bool isInt(char *pVal, uint16_t len) {
1278 1279
  if (len <= 3) {
    return false;
1280
  }
1281
  if (strcasecmp(&pVal[len - 3], "i32") == 0) {
1282
    //printf("Type is int32(%s)\n", pVal);
1283 1284 1285
    return true;
  }
  return false;
1286 1287
}

1288
static bool isUint(char *pVal, uint16_t len) {
1289 1290 1291 1292 1293 1294
  if (len <= 3) {
    return false;
  }
  if (pVal[0] == '-') {
    return false;
  }
1295
  if (strcasecmp(&pVal[len - 3], "u32") == 0) {
1296
    //printf("Type is uint32(%s)\n", pVal);
1297 1298 1299
    return true;
  }
  return false;
1300 1301
}

1302
static bool isBigInt(char *pVal, uint16_t len) {
1303 1304
  if (len <= 3) {
    return false;
1305
  }
1306
  if (strcasecmp(&pVal[len - 3], "i64") == 0) {
1307
    //printf("Type is int64(%s)\n", pVal);
1308 1309 1310
    return true;
  }
  return false;
1311 1312
}

1313
static bool isBigUint(char *pVal, uint16_t len) {
1314 1315 1316 1317 1318
  if (len <= 3) {
    return false;
  }
  if (pVal[0] == '-') {
    return false;
1319
  }
1320
  if (strcasecmp(&pVal[len - 3], "u64") == 0) {
1321
    //printf("Type is uint64(%s)\n", pVal);
1322 1323 1324
    return true;
  }
  return false;
1325 1326
}

1327
static bool isFloat(char *pVal, uint16_t len) {
1328 1329 1330
  if (len <= 3) {
    return false;
  }
1331
  if (strcasecmp(&pVal[len - 3], "f32") == 0) {
1332
    //printf("Type is float(%s)\n", pVal);
1333 1334 1335
    return true;
  }
  return false;
1336 1337
}

1338
static bool isDouble(char *pVal, uint16_t len) {
1339 1340 1341
  if (len <= 3) {
    return false;
  }
1342
  if (strcasecmp(&pVal[len - 3], "f64") == 0) {
1343
    //printf("Type is double(%s)\n", pVal);
1344 1345 1346 1347 1348
    return true;
  }
  return false;
}

1349
static bool isBool(char *pVal, uint16_t len, bool *bVal) {
1350
  if ((len == 1) && !strcasecmp(&pVal[len - 1], "t")) {
1351 1352
    //printf("Type is bool(%c)\n", pVal[len - 1]);
    *bVal = true;
1353
    return true;
1354
  }
1355

1356
  if ((len == 1) && !strcasecmp(&pVal[len - 1], "f")) {
1357 1358
    //printf("Type is bool(%c)\n", pVal[len - 1]);
    *bVal = false;
1359
    return true;
1360
  }
1361

1362
  if((len == 4) && !strcasecmp(&pVal[len - 4], "true")) {
1363 1364
    //printf("Type is bool(%s)\n", &pVal[len - 4]);
    *bVal = true;
1365 1366
    return true;
  }
1367
  if((len == 5) && !strcasecmp(&pVal[len - 5], "false")) {
1368 1369
    //printf("Type is bool(%s)\n", &pVal[len - 5]);
    *bVal = false;
1370 1371 1372
    return true;
  }
  return false;
1373 1374
}

1375
static bool isBinary(char *pVal, uint16_t len) {
1376 1377 1378 1379 1380 1381
  //binary: "abc"
  if (len < 2) {
    return false;
  }
  //binary
  if (pVal[0] == '"' && pVal[len - 1] == '"') {
1382
    //printf("Type is binary(%s)\n", pVal);
1383 1384 1385 1386
    return true;
  }
  return false;
}
1387

1388
static bool isNchar(char *pVal, uint16_t len) {
1389 1390
  //nchar: L"abc"
  if (len < 3) {
1391 1392
    return false;
  }
1393
  if ((pVal[0] == 'l' || pVal[0] == 'L')&& pVal[1] == '"' && pVal[len - 1] == '"') {
1394
    //printf("Type is nchar(%s)\n", pVal);
1395
    return true;
1396
  }
1397 1398 1399
  return false;
}

1400
static bool isTimeStamp(char *pVal, uint16_t len, SMLTimeStampType *tsType) {
1401 1402 1403 1404 1405
  if (len == 0) {
    return true;
  }
  if ((len == 1) && pVal[0] == '0') {
    *tsType = SML_TIME_STAMP_NOW;
1406
    //printf("Type is timestamp(%s)\n", pVal);
1407 1408 1409 1410 1411 1412 1413 1414
    return true;
  }
  if (len < 2) {
    return false;
  }
  //No appendix use usec as default
  if (isdigit(pVal[len - 1]) && isdigit(pVal[len - 2])) {
    *tsType = SML_TIME_STAMP_MICRO_SECONDS;
1415
    //printf("Type is timestamp(%s)\n", pVal);
1416 1417 1418 1419 1420 1421 1422 1423 1424 1425 1426 1427 1428 1429 1430 1431
    return true;
  }
  if (pVal[len - 1] == 's') {
    switch (pVal[len - 2]) {
      case 'm':
        *tsType = SML_TIME_STAMP_MILLI_SECONDS;
        break;
      case 'u':
        *tsType = SML_TIME_STAMP_MICRO_SECONDS;
        break;
      case 'n':
        *tsType = SML_TIME_STAMP_NANO_SECONDS;
        break;
      default:
        if (isdigit(pVal[len - 2])) {
          *tsType = SML_TIME_STAMP_SECONDS;
1432
          break;
1433
        } else {
1434 1435 1436
          return false;
        }
    }
1437
    //printf("Type is timestamp(%s)\n", pVal);
1438 1439 1440 1441
    return true;
  }
  return false;
}
1442

1443
static bool convertStrToNumber(TAOS_SML_KV *pVal, char *str, SSmlLinesInfo* info) {
1444 1445 1446 1447 1448 1449 1450
  errno = 0;
  uint8_t type = pVal->type;
  int16_t length = pVal->length;
  int64_t val_s;
  uint64_t val_u;
  double val_d;

1451
  strntolower_s(str, str, (int32_t)strlen(str));
1452 1453 1454 1455 1456 1457 1458 1459 1460 1461 1462
  if (IS_FLOAT_TYPE(type)) {
    val_d = strtod(str, NULL);
  } else {
    if (IS_SIGNED_NUMERIC_TYPE(type)) {
      val_s = strtoll(str, NULL, 10);
    } else {
      val_u = strtoull(str, NULL, 10);
    }
  }

  if (errno == ERANGE) {
1463
    tscError("SML:0x%"PRIx64" Convert number(%s) out of range", info->id, str);
1464 1465 1466 1467 1468 1469 1470 1471 1472 1473 1474 1475 1476 1477 1478 1479 1480 1481 1482 1483 1484 1485 1486 1487 1488 1489 1490 1491 1492 1493 1494 1495 1496 1497 1498 1499 1500 1501 1502 1503 1504 1505 1506 1507 1508 1509 1510 1511 1512 1513 1514 1515 1516 1517 1518 1519 1520 1521 1522 1523 1524 1525 1526 1527 1528 1529 1530 1531 1532 1533 1534 1535 1536 1537 1538 1539 1540 1541 1542
    return false;
  }

  switch (type) {
    case TSDB_DATA_TYPE_TINYINT:
      if (!IS_VALID_TINYINT(val_s)) {
        return false;
      }
      pVal->value = calloc(length, 1);
      *(int8_t *)(pVal->value) = (int8_t)val_s;
      break;
    case TSDB_DATA_TYPE_UTINYINT:
      if (!IS_VALID_UTINYINT(val_u)) {
        return false;
      }
      pVal->value = calloc(length, 1);
      *(uint8_t *)(pVal->value) = (uint8_t)val_u;
      break;
    case TSDB_DATA_TYPE_SMALLINT:
      if (!IS_VALID_SMALLINT(val_s)) {
        return false;
      }
      pVal->value = calloc(length, 1);
      *(int16_t *)(pVal->value) = (int16_t)val_s;
      break;
    case TSDB_DATA_TYPE_USMALLINT:
      if (!IS_VALID_USMALLINT(val_u)) {
        return false;
      }
      pVal->value = calloc(length, 1);
      *(uint16_t *)(pVal->value) = (uint16_t)val_u;
      break;
    case TSDB_DATA_TYPE_INT:
      if (!IS_VALID_INT(val_s)) {
        return false;
      }
      pVal->value = calloc(length, 1);
      *(int32_t *)(pVal->value) = (int32_t)val_s;
      break;
    case TSDB_DATA_TYPE_UINT:
      if (!IS_VALID_UINT(val_u)) {
        return false;
      }
      pVal->value = calloc(length, 1);
      *(uint32_t *)(pVal->value) = (uint32_t)val_u;
      break;
    case TSDB_DATA_TYPE_BIGINT:
      if (!IS_VALID_BIGINT(val_s)) {
        return false;
      }
      pVal->value = calloc(length, 1);
      *(int64_t *)(pVal->value) = (int64_t)val_s;
      break;
    case TSDB_DATA_TYPE_UBIGINT:
      if (!IS_VALID_UBIGINT(val_u)) {
        return false;
      }
      pVal->value = calloc(length, 1);
      *(uint64_t *)(pVal->value) = (uint64_t)val_u;
      break;
    case TSDB_DATA_TYPE_FLOAT:
      if (!IS_VALID_FLOAT(val_d)) {
        return false;
      }
      pVal->value = calloc(length, 1);
      *(float *)(pVal->value) = (float)val_d;
      break;
    case TSDB_DATA_TYPE_DOUBLE:
      if (!IS_VALID_DOUBLE(val_d)) {
        return false;
      }
      pVal->value = calloc(length, 1);
      *(double *)(pVal->value) = (double)val_d;
      break;
    default:
      return false;
  }
  return true;
}
1543
//len does not include '\0' from value.
1544
bool convertSmlValueType(TAOS_SML_KV *pVal, char *value,
1545
                         uint16_t len, SSmlLinesInfo* info, bool isTag) {
1546 1547 1548
  if (len <= 0) {
    return false;
  }
G
Ganlin Zhao 已提交
1549

1550 1551 1552 1553 1554 1555 1556 1557 1558
  //convert tags value to Nchar
  if (isTag) {
    pVal->type = TSDB_DATA_TYPE_NCHAR;
    pVal->length = len;
    pVal->value = calloc(pVal->length, 1);
    memcpy(pVal->value, value, pVal->length);
    return true;
  }

1559
  //integer number
1560 1561 1562 1563 1564 1565 1566 1567 1568 1569
  bool has_sign;
  if (isInteger(value, len, &has_sign)) {
    pVal->type = has_sign ? TSDB_DATA_TYPE_BIGINT : TSDB_DATA_TYPE_UBIGINT;
    pVal->length = (int16_t)tDataTypes[pVal->type].bytes;
    value[len - 1] = '\0';
    if (!isValidInteger(value) || !convertStrToNumber(pVal, value, info)) {
      return false;
    }
    return true;
  }
1570
  if (isTinyInt(value, len)) {
1571 1572 1573
    pVal->type = TSDB_DATA_TYPE_TINYINT;
    pVal->length = (int16_t)tDataTypes[pVal->type].bytes;
    value[len - 2] = '\0';
1574
    if (!isValidInteger(value) || !convertStrToNumber(pVal, value, info)) {
1575
      return false;
1576 1577 1578
    }
    return true;
  }
1579
  if (isTinyUint(value, len)) {
1580 1581 1582
    pVal->type = TSDB_DATA_TYPE_UTINYINT;
    pVal->length = (int16_t)tDataTypes[pVal->type].bytes;
    value[len - 2] = '\0';
1583
    if (!isValidInteger(value) || !convertStrToNumber(pVal, value, info)) {
1584 1585 1586 1587
      return false;
    }
    return true;
  }
1588
  if (isSmallInt(value, len)) {
1589 1590 1591
    pVal->type = TSDB_DATA_TYPE_SMALLINT;
    pVal->length = (int16_t)tDataTypes[pVal->type].bytes;
    value[len - 3] = '\0';
1592
    if (!isValidInteger(value) || !convertStrToNumber(pVal, value, info)) {
1593 1594 1595 1596
      return false;
    }
    return true;
  }
1597
  if (isSmallUint(value, len)) {
1598 1599 1600
    pVal->type = TSDB_DATA_TYPE_USMALLINT;
    pVal->length = (int16_t)tDataTypes[pVal->type].bytes;
    value[len - 3] = '\0';
1601
    if (!isValidInteger(value) || !convertStrToNumber(pVal, value, info)) {
1602 1603 1604 1605
      return false;
    }
    return true;
  }
1606
  if (isInt(value, len)) {
1607 1608 1609
    pVal->type = TSDB_DATA_TYPE_INT;
    pVal->length = (int16_t)tDataTypes[pVal->type].bytes;
    value[len - 3] = '\0';
1610
    if (!isValidInteger(value) || !convertStrToNumber(pVal, value, info)) {
1611 1612 1613 1614
      return false;
    }
    return true;
  }
1615
  if (isUint(value, len)) {
1616 1617 1618
    pVal->type = TSDB_DATA_TYPE_UINT;
    pVal->length = (int16_t)tDataTypes[pVal->type].bytes;
    value[len - 3] = '\0';
1619
    if (!isValidInteger(value) || !convertStrToNumber(pVal, value, info)) {
1620 1621 1622 1623
      return false;
    }
    return true;
  }
1624
  if (isBigInt(value, len)) {
1625 1626 1627
    pVal->type = TSDB_DATA_TYPE_BIGINT;
    pVal->length = (int16_t)tDataTypes[pVal->type].bytes;
    value[len - 3] = '\0';
1628
    if (!isValidInteger(value) || !convertStrToNumber(pVal, value, info)) {
1629 1630 1631 1632
      return false;
    }
    return true;
  }
1633
  if (isBigUint(value, len)) {
1634 1635 1636
    pVal->type = TSDB_DATA_TYPE_UBIGINT;
    pVal->length = (int16_t)tDataTypes[pVal->type].bytes;
    value[len - 3] = '\0';
1637
    if (!isValidInteger(value) || !convertStrToNumber(pVal, value, info)) {
1638 1639 1640 1641
      return false;
    }
    return true;
  }
1642 1643 1644 1645 1646
  //floating number
  if (isFloat(value, len)) {
    pVal->type = TSDB_DATA_TYPE_FLOAT;
    pVal->length = (int16_t)tDataTypes[pVal->type].bytes;
    value[len - 3] = '\0';
1647
    if (!isValidFloat(value) || !convertStrToNumber(pVal, value, info)) {
1648 1649 1650 1651 1652 1653 1654 1655
      return false;
    }
    return true;
  }
  if (isDouble(value, len)) {
    pVal->type = TSDB_DATA_TYPE_DOUBLE;
    pVal->length = (int16_t)tDataTypes[pVal->type].bytes;
    value[len - 3] = '\0';
1656
    if (!isValidFloat(value) || !convertStrToNumber(pVal, value, info)) {
1657 1658 1659 1660 1661 1662 1663 1664 1665 1666 1667 1668 1669 1670 1671 1672 1673 1674 1675 1676 1677 1678 1679 1680 1681 1682 1683 1684 1685 1686 1687
      return false;
    }
    return true;
  }
  //binary
  if (isBinary(value, len)) {
    pVal->type = TSDB_DATA_TYPE_BINARY;
    pVal->length = len - 2;
    pVal->value = calloc(pVal->length, 1);
    //copy after "
    memcpy(pVal->value, value + 1, pVal->length);
    return true;
  }
  //nchar
  if (isNchar(value, len)) {
    pVal->type = TSDB_DATA_TYPE_NCHAR;
    pVal->length = len - 3;
    pVal->value = calloc(pVal->length, 1);
    //copy after L"
    memcpy(pVal->value, value + 2, pVal->length);
    return true;
  }
  //bool
  bool bVal;
  if (isBool(value, len, &bVal)) {
    pVal->type = TSDB_DATA_TYPE_BOOL;
    pVal->length = (int16_t)tDataTypes[pVal->type].bytes;
    pVal->value = calloc(pVal->length, 1);
    memcpy(pVal->value, &bVal, pVal->length);
    return true;
  }
1688

1689 1690
  //Handle default(no appendix) type as DOUBLE
  if (isValidInteger(value) || isValidFloat(value)) {
1691
    pVal->type = TSDB_DATA_TYPE_DOUBLE;
G
Ganlin Zhao 已提交
1692
    pVal->length = (int16_t)tDataTypes[pVal->type].bytes;
1693
    if (!convertStrToNumber(pVal, value, info)) {
1694 1695
      return false;
    }
G
Ganlin Zhao 已提交
1696 1697
    return true;
  }
1698
  return false;
1699
}
1700

1701 1702
static int32_t getTimeStampValue(char *value, uint16_t len,
                                 SMLTimeStampType type, int64_t *ts) {
1703 1704 1705 1706

  if (len >= 2) {
    for (int i = 0; i < len - 2; ++i) {
      if(!isdigit(value[i])) {
1707
        return TSDB_CODE_TSC_INVALID_TIME_STAMP;
1708
      }
1709 1710 1711 1712 1713 1714 1715 1716 1717 1718 1719 1720 1721
    }
  }
  //No appendix or no timestamp given (len = 0)
  if (len >= 1 && isdigit(value[len - 1]) && type != SML_TIME_STAMP_NOW) {
    type = SML_TIME_STAMP_MICRO_SECONDS;
  }
  if (len != 0) {
    *ts = (int64_t)strtoll(value, NULL, 10);
  } else {
    type = SML_TIME_STAMP_NOW;
  }
  switch (type) {
    case SML_TIME_STAMP_NOW: {
1722
      *ts = taosGetTimestampNs();
1723
      break;
1724 1725
    }
    case SML_TIME_STAMP_SECONDS: {
1726
      *ts = (int64_t)(*ts * 1e9);
1727
      break;
1728 1729
    }
    case SML_TIME_STAMP_MILLI_SECONDS: {
1730
      *ts = convertTimePrecision(*ts, TSDB_TIME_PRECISION_MILLI, TSDB_TIME_PRECISION_NANO);
1731
      break;
1732 1733
    }
    case SML_TIME_STAMP_MICRO_SECONDS: {
1734
      *ts = convertTimePrecision(*ts, TSDB_TIME_PRECISION_MICRO, TSDB_TIME_PRECISION_NANO);
1735
      break;
1736 1737
    }
    case SML_TIME_STAMP_NANO_SECONDS: {
1738
      *ts = *ts * 1;
1739 1740 1741
      break;
    }
    default: {
1742
      return TSDB_CODE_TSC_INVALID_TIME_STAMP;
1743
    }
1744
  }
1745
  return TSDB_CODE_SUCCESS;
1746 1747
}

1748 1749
int32_t convertSmlTimeStamp(TAOS_SML_KV *pVal, char *value,
                            uint16_t len, SSmlLinesInfo* info) {
1750 1751 1752
  int32_t ret;
  SMLTimeStampType type;
  int64_t tsVal;
1753

1754
  strntolower_s(value, value, len);
1755
  if (!isTimeStamp(value, len, &type)) {
1756
    return TSDB_CODE_TSC_INVALID_TIME_STAMP;
1757 1758
  }

1759
  ret = getTimeStampValue(value, len, type, &tsVal);
1760 1761 1762
  if (ret) {
    return ret;
  }
1763
  tscDebug("SML:0x%"PRIx64"Timestamp after conversion:%"PRId64, info->id, tsVal);
1764 1765 1766 1767 1768 1769 1770 1771

  pVal->type = TSDB_DATA_TYPE_TIMESTAMP;
  pVal->length = (int16_t)tDataTypes[pVal->type].bytes;
  pVal->value = calloc(pVal->length, 1);
  memcpy(pVal->value, &tsVal, pVal->length);
  return TSDB_CODE_SUCCESS;
}

1772
static int32_t parseSmlTimeStamp(TAOS_SML_KV **pTS, const char **index, SSmlLinesInfo* info) {
1773
  const char *start, *cur;
1774 1775 1776
  int32_t ret = TSDB_CODE_SUCCESS;
  int len = 0;
  char key[] = "_ts";
1777
  char *value = NULL;
1778

1779
  start = cur = *index;
1780
  *pTS = calloc(1, sizeof(TAOS_SML_KV));
1781

1782
  while(*cur != '\0') {
1783 1784 1785 1786
    cur++;
    len++;
  }

1787
  if (len > 0) {
1788
    value = calloc(len + 1, 1);
1789 1790 1791
    memcpy(value, start, len);
  }

1792
  ret = convertSmlTimeStamp(*pTS, value, len, info);
1793
  if (ret) {
1794
    free(value);
1795 1796
    free(*pTS);
    return ret;
1797
  }
1798
  free(value);
1799

1800 1801 1802
  (*pTS)->key = calloc(sizeof(key), 1);
  memcpy((*pTS)->key, key, sizeof(key));
  return ret;
1803
}
1804

1805
bool checkDuplicateKey(char *key, SHashObj *pHash, SSmlLinesInfo* info) {
1806 1807 1808 1809 1810 1811 1812 1813 1814 1815 1816 1817 1818
  char *val = NULL;
  char *cur = key;
  char keyLower[TSDB_COL_NAME_LEN];
  size_t keyLen = 0;
  while(*cur != '\0') {
    keyLower[keyLen] = tolower(*cur);
    keyLen++;
    cur++;
  }
  keyLower[keyLen] = '\0';

  val = taosHashGet(pHash, keyLower, keyLen);
  if (val) {
1819
    tscError("SML:0x%"PRIx64" Duplicate key detected:%s", info->id, keyLower);
1820 1821 1822 1823 1824 1825 1826 1827 1828
    return true;
  }

  uint8_t dummy_val = 0;
  taosHashPut(pHash, keyLower, strlen(key), &dummy_val, sizeof(uint8_t));

  return false;
}

1829
static int32_t parseSmlKey(TAOS_SML_KV *pKV, const char **index, SHashObj *pHash, SSmlLinesInfo* info) {
1830
  const char *cur = *index;
1831
  char key[TSDB_COL_NAME_LEN + 1];  // +1 to avoid key[len] over write
1832 1833
  uint16_t len = 0;

G
Ganlin Zhao 已提交
1834 1835
  //key field cannot start with digit
  if (isdigit(*cur)) {
1836
    tscError("SML:0x%"PRIx64" Tag key cannnot start with digit", info->id);
1837
    return TSDB_CODE_TSC_LINE_SYNTAX_ERROR;
1838 1839
  }
  while (*cur != '\0') {
1840 1841
    if (len >= TSDB_COL_NAME_LEN - 1) {
      tscError("SML:0x%"PRIx64" Key field cannot exceeds %d characters", info->id, TSDB_COL_NAME_LEN - 1);
1842
      return TSDB_CODE_TSC_INVALID_COLUMN_LENGTH;
1843 1844 1845 1846 1847 1848 1849
    }
    //unescaped '=' identifies a tag key
    if (*cur == '=' && *(cur - 1) != '\\') {
      break;
    }
    //Escape special character
    if (*cur == '\\') {
1850
      escapeSpecialCharacter(2, &cur);
1851
    }
1852 1853 1854 1855 1856
    key[len] = *cur;
    cur++;
    len++;
  }
  key[len] = '\0';
1857

1858
  if (checkDuplicateKey(key, pHash, info)) {
1859 1860 1861
    return TSDB_CODE_TSC_LINE_SYNTAX_ERROR;
  }

1862 1863
  pKV->key = calloc(len + 1, 1);
  memcpy(pKV->key, key, len + 1);
1864
  //tscDebug("SML:0x%"PRIx64" Key:%s|len:%d", info->id, pKV->key, len);
1865
  *index = cur + 1;
1866
  return TSDB_CODE_SUCCESS;
1867
}
1868

1869

1870
static bool parseSmlValue(TAOS_SML_KV *pKV, const char **index,
1871
                          bool *is_last_kv, SSmlLinesInfo* info, bool isTag) {
1872 1873
  const char *start, *cur;
  char *value = NULL;
1874
  uint16_t len = 0;
1875 1876
  start = cur = *index;

1877 1878 1879 1880 1881 1882 1883 1884 1885 1886 1887
  while (1) {
    // unescaped ',' or ' ' or '\0' identifies a value
    if ((*cur == ',' || *cur == ' ' || *cur == '\0') && *(cur - 1) != '\\') {
      //unescaped ' ' or '\0' indicates end of value
      *is_last_kv = (*cur == ' ' || *cur == '\0') ? true : false;
      break;
    }
    //Escape special character
    if (*cur == '\\') {
      escapeSpecialCharacter(2, &cur);
    }
1888 1889 1890
    cur++;
    len++;
  }
1891

1892 1893 1894
  value = calloc(len + 1, 1);
  memcpy(value, start, len);
  value[len] = '\0';
1895
  if (!convertSmlValueType(pKV, value, len, info, isTag)) {
1896 1897
    tscError("SML:0x%"PRIx64" Failed to convert sml value string(%s) to any type",
            info->id, value);
1898 1899
    //free previous alocated key field
    free(pKV->key);
1900
    pKV->key = NULL;
1901
    free(value);
1902
    return TSDB_CODE_TSC_INVALID_VALUE;
1903
  }
1904
  free(value);
1905

1906 1907 1908 1909 1910
  *index = (*cur == '\0') ? cur : cur + 1;
  return TSDB_CODE_SUCCESS;
}

static int32_t parseSmlMeasurement(TAOS_SML_DATA_POINT *pSml, const char **index,
1911
                                   uint8_t *has_tags, SSmlLinesInfo* info) {
1912 1913 1914
  const char *cur = *index;
  uint16_t len = 0;

1915 1916 1917 1918
  pSml->stableName = calloc(TSDB_TABLE_NAME_LEN + 1, 1);    // +1 to avoid 1772 line over write
  if (pSml->stableName == NULL){
      return TSDB_CODE_TSC_OUT_OF_MEMORY;
  }
G
Ganlin Zhao 已提交
1919
  if (isdigit(*cur)) {
1920
    tscError("SML:0x%"PRIx64" Measurement field cannnot start with digit", info->id);
1921
    free(pSml->stableName);
1922
    pSml->stableName = NULL;
1923
    return TSDB_CODE_TSC_LINE_SYNTAX_ERROR;
1924 1925
  }

1926
  while (*cur != '\0') {
1927 1928
    if (len >= TSDB_TABLE_NAME_LEN - 1) {
      tscError("SML:0x%"PRIx64" Measurement field cannot exceeds %d characters", info->id, TSDB_TABLE_NAME_LEN - 1);
1929
      free(pSml->stableName);
1930
      pSml->stableName = NULL;
1931
      return TSDB_CODE_TSC_INVALID_TABLE_ID_LENGTH;
1932 1933 1934 1935 1936 1937 1938 1939 1940 1941 1942 1943 1944 1945
    }
    //first unescaped comma or space identifies measurement
    //if space detected first, meaning no tag in the input
    if (*cur == ',' && *(cur - 1) != '\\') {
      *has_tags = 1;
      break;
    }
    if (*cur == ' ' && *(cur - 1) != '\\') {
      break;
    }
    //Comma, Space, Backslash needs to be escaped if any
    if (*cur == '\\') {
      escapeSpecialCharacter(1, &cur);
    }
1946
    pSml->stableName[len] = tolower(*cur);
1947 1948 1949 1950 1951
    cur++;
    len++;
  }
  pSml->stableName[len] = '\0';
  *index = cur + 1;
1952
  tscDebug("SML:0x%"PRIx64" Stable name in measurement:%s|len:%d", info->id, pSml->stableName, len);
1953 1954

  return TSDB_CODE_SUCCESS;
1955
}
1956

1957
//Table name can only contain digits(0-9),alphebet(a-z),underscore(_)
1958 1959
int32_t isValidChildTableName(const char *pTbName, int16_t len, SSmlLinesInfo* info) {
  if (len > TSDB_TABLE_NAME_LEN - 1) {
1960
    tscError("SML:0x%"PRIx64" child table name cannot exceeds %d characters", info->id, TSDB_TABLE_NAME_LEN - 1);
1961 1962
    return TSDB_CODE_TSC_INVALID_TABLE_ID_LENGTH;
  }
1963 1964
  const char *cur = pTbName;
  for (int i = 0; i < len; ++i) {
1965
    if(!isdigit(cur[i]) && !isalpha(cur[i]) && (cur[i] != '_')) {
1966 1967 1968 1969 1970 1971 1972
      return TSDB_CODE_TSC_LINE_SYNTAX_ERROR;
    }
  }
  return TSDB_CODE_SUCCESS;
}


1973
static int32_t parseSmlKvPairs(TAOS_SML_KV **pKVs, int *num_kvs,
1974
                               const char **index, bool isField,
1975 1976
                               TAOS_SML_DATA_POINT* smlData, SHashObj *pHash,
                               SSmlLinesInfo* info) {
1977
  const char *cur = *index;
1978
  int32_t ret = TSDB_CODE_SUCCESS;
1979 1980
  TAOS_SML_KV *pkv;
  bool is_last_kv = false;
1981

1982
  int32_t capacity = 0;
1983
  if (isField) {
1984 1985 1986
    capacity = 64;
    *pKVs = calloc(capacity, sizeof(TAOS_SML_KV));
    // leave space for timestamp;
1987 1988
    pkv = *pKVs;
    pkv++;
1989 1990 1991
  } else {
    capacity = 8;
    *pKVs = calloc(capacity, sizeof(TAOS_SML_KV));
1992 1993
    pkv = *pKVs;
  }
1994

1995
  while (*cur != '\0') {
1996
    ret = parseSmlKey(pkv, &cur, pHash, info);
1997
    if (ret) {
1998
      tscError("SML:0x%"PRIx64" Unable to parse key", info->id);
1999 2000
      goto error;
    }
2001
    ret = parseSmlValue(pkv, &cur, &is_last_kv, info, !isField);
2002
    if (ret) {
2003
      tscError("SML:0x%"PRIx64" Unable to parse value", info->id);
2004 2005
      goto error;
    }
2006 2007
    if (!isField &&
        (strcasecmp(pkv->key, "ID") == 0) && pkv->type == TSDB_DATA_TYPE_BINARY) {
2008
      ret = isValidChildTableName(pkv->value, pkv->length, info);
2009
      if (ret) {
2010 2011
        goto error;
      }
2012 2013
      smlData->childTableName = malloc( pkv->length + 1);
      memcpy(smlData->childTableName, pkv->value, pkv->length);
2014
      strntolower_s(smlData->childTableName, smlData->childTableName, (int32_t)pkv->length);
2015 2016 2017 2018 2019 2020
      smlData->childTableName[pkv->length] = '\0';
      free(pkv->key);
      free(pkv->value);
    } else {
      *num_kvs += 1;
    }
2021
    if (is_last_kv) {
2022 2023 2024 2025 2026
      goto done;
    }

    //reallocate addtional memory for more kvs
    TAOS_SML_KV *more_kvs = NULL;
2027

2028
    if (isField) {
2029 2030
      if ((*num_kvs + 2) > capacity) {
        capacity *= 3; capacity /= 2;
2031 2032 2033
        more_kvs = realloc(*pKVs, capacity * sizeof(TAOS_SML_KV));
      } else {
        more_kvs = *pKVs;
2034
      }
2035
    } else {
2036 2037
      if ((*num_kvs + 1) > capacity) {
        capacity *= 3; capacity /= 2;
2038 2039 2040
        more_kvs = realloc(*pKVs, capacity * sizeof(TAOS_SML_KV));
      } else {
        more_kvs = *pKVs;
2041
      }
2042
    }
2043

2044
    if (!more_kvs) {
2045 2046
      goto error;
    }
2047 2048 2049 2050 2051 2052 2053 2054 2055
    *pKVs = more_kvs;
    //move pKV points to next TAOS_SML_KV block
    if (isField) {
      pkv = *pKVs + *num_kvs + 1;
    } else {
      pkv = *pKVs + *num_kvs;
    }
  }
  goto done;
2056

2057
error:
2058
  return ret;
2059
done:
2060
  *index = cur;
2061
  return ret;
2062
}
2063

2064 2065 2066 2067 2068 2069 2070 2071 2072 2073 2074 2075 2076
static void moveTimeStampToFirstKv(TAOS_SML_DATA_POINT** smlData, TAOS_SML_KV *ts) {
  TAOS_SML_KV* tsField = (*smlData)->fields;
  tsField->length = ts->length;
  tsField->type = ts->type;
  tsField->value = malloc(ts->length);
  tsField->key = malloc(strlen(ts->key) + 1);
  memcpy(tsField->key, ts->key, strlen(ts->key) + 1);
  memcpy(tsField->value, ts->value, ts->length);
  (*smlData)->fieldNum = (*smlData)->fieldNum + 1;

  free(ts->key);
  free(ts->value);
  free(ts);
2077 2078
}

2079
int32_t tscParseLine(const char* sql, TAOS_SML_DATA_POINT* smlData, SSmlLinesInfo* info) {
2080
  const char* index = sql;
2081
  int32_t ret = TSDB_CODE_SUCCESS;
2082 2083
  uint8_t has_tags = 0;
  TAOS_SML_KV *timestamp = NULL;
2084
  SHashObj *keyHashTable = taosHashInit(128, taosGetDefaultHashFunction(TSDB_DATA_TYPE_BINARY), true, false);
2085

2086
  ret = parseSmlMeasurement(smlData, &index, &has_tags, info);
2087
  if (ret) {
2088
    tscError("SML:0x%"PRIx64" Unable to parse measurement", info->id);
2089
    taosHashCleanup(keyHashTable);
2090
    return ret;
2091
  }
2092
  tscDebug("SML:0x%"PRIx64" Parse measurement finished, has_tags:%d", info->id, has_tags);
2093 2094 2095

  //Parse Tags
  if (has_tags) {
2096
    ret = parseSmlKvPairs(&smlData->tags, &smlData->tagNum, &index, false, smlData, keyHashTable, info);
2097
    if (ret) {
2098
      tscError("SML:0x%"PRIx64" Unable to parse tag", info->id);
2099
      taosHashCleanup(keyHashTable);
2100 2101
      return ret;
    }
2102
  }
2103
  tscDebug("SML:0x%"PRIx64" Parse tags finished, num of tags:%d", info->id, smlData->tagNum);
2104 2105

  //Parse fields
2106
  ret = parseSmlKvPairs(&smlData->fields, &smlData->fieldNum, &index, true, smlData, keyHashTable, info);
2107
  if (ret) {
2108
    tscError("SML:0x%"PRIx64" Unable to parse field", info->id);
2109
    taosHashCleanup(keyHashTable);
2110
    return ret;
2111
  }
2112
  tscDebug("SML:0x%"PRIx64" Parse fields finished, num of fields:%d", info->id, smlData->fieldNum);
2113
  taosHashCleanup(keyHashTable);
2114

2115
  //Parse timestamp
2116
  ret = parseSmlTimeStamp(&timestamp, &index, info);
2117
  if (ret) {
2118
    tscError("SML:0x%"PRIx64" Unable to parse timestamp", info->id);
2119
    return ret;
2120
  }
2121
  moveTimeStampToFirstKv(&smlData, timestamp);
2122
  tscDebug("SML:0x%"PRIx64" Parse timestamp finished", info->id);
2123

2124
  return TSDB_CODE_SUCCESS;
2125 2126
}

2127
//=========================================================================
2128

S
shenglian zhou 已提交
2129 2130 2131 2132 2133 2134 2135 2136 2137 2138 2139 2140 2141 2142 2143
void destroySmlDataPoint(TAOS_SML_DATA_POINT* point) {
  for (int i=0; i<point->tagNum; ++i) {
    free((point->tags+i)->key);
    free((point->tags+i)->value);
  }
  free(point->tags);
  for (int i=0; i<point->fieldNum; ++i) {
    free((point->fields+i)->key);
    free((point->fields+i)->value);
  }
  free(point->fields);
  free(point->stableName);
  free(point->childTableName);
}

S
shenglian zhou 已提交
2144
int32_t tscParseLines(char* lines[], int numLines, SArray* points, SArray* failedLines, SSmlLinesInfo* info) {
2145
  for (int32_t i = 0; i < numLines; ++i) {
2146
    TAOS_SML_DATA_POINT point = {0};
2147
    int32_t code = tscParseLine(lines[i], &point, info);
2148
    if (code != TSDB_CODE_SUCCESS) {
S
shenglian zhou 已提交
2149
      tscError("SML:0x%"PRIx64" data point line parse failed. line %d : %s", info->id, i, lines[i]);
S
shenglian zhou 已提交
2150
      destroySmlDataPoint(&point);
2151
      return code;
2152
    } else {
S
shenglian zhou 已提交
2153
      tscDebug("SML:0x%"PRIx64" data point line parse success. line %d", info->id, i);
2154 2155
    }

2156 2157
    taosArrayPush(points, &point);
  }
2158
  return TSDB_CODE_SUCCESS;
2159 2160
}

2161
int taos_insert_lines(TAOS* taos, char* lines[], int numLines) {
2162
  int32_t code = 0;
2163

2164
  SSmlLinesInfo* info = tcalloc(1, sizeof(SSmlLinesInfo));
S
shenglian zhou 已提交
2165 2166
  info->id = genLinesSmlId();

2167
  if (numLines <= 0 || numLines > 65536) {
S
shenglian zhou 已提交
2168
    tscError("SML:0x%"PRIx64" taos_insert_lines numLines should be between 1 and 65536. numLines: %d", info->id, numLines);
2169
    tfree(info);
2170 2171 2172 2173 2174 2175
    code = TSDB_CODE_TSC_APP_ERROR;
    return code;
  }

  for (int i = 0; i < numLines; ++i) {
    if (lines[i] == NULL) {
S
shenglian zhou 已提交
2176
      tscError("SML:0x%"PRIx64" taos_insert_lines line %d is NULL", info->id, i);
2177
      tfree(info);
2178 2179 2180 2181 2182
      code = TSDB_CODE_TSC_APP_ERROR;
      return code;
    }
  }

2183
  SArray* lpPoints = taosArrayInit(numLines, sizeof(TAOS_SML_DATA_POINT));
2184
  if (lpPoints == NULL) {
S
shenglian zhou 已提交
2185
    tscError("SML:0x%"PRIx64" taos_insert_lines failed to allocate memory", info->id);
2186
    tfree(info);
2187 2188
    return TSDB_CODE_TSC_OUT_OF_MEMORY;
  }
2189

S
shenglian zhou 已提交
2190 2191
  tscDebug("SML:0x%"PRIx64" taos_insert_lines begin inserting %d lines, first line: %s", info->id, numLines, lines[0]);
  code = tscParseLines(lines, numLines, lpPoints, NULL, info);
S
shenglian zhou 已提交
2192 2193
  size_t numPoints = taosArrayGetSize(lpPoints);

2194 2195
  if (code != 0) {
    goto cleanup;
2196 2197
  }

2198
  TAOS_SML_DATA_POINT* points = TARRAY_GET_START(lpPoints);
2199
  code = tscSmlInsert(taos, points, (int)numPoints, info);
2200
  if (code != 0) {
S
shenglian zhou 已提交
2201
    tscError("SML:0x%"PRIx64" taos_sml_insert error: %s", info->id, tstrerror((code)));
2202
  }
S
Shenglian Zhou 已提交
2203

2204
cleanup:
S
shenglian zhou 已提交
2205
  tscDebug("SML:0x%"PRIx64" taos_insert_lines finish inserting %d lines. code: %d", info->id, numLines, code);
2206 2207
  points = TARRAY_GET_START(lpPoints);
  numPoints = taosArrayGetSize(lpPoints);
S
Shenglian Zhou 已提交
2208 2209 2210
  for (int i=0; i<numPoints; ++i) {
    destroySmlDataPoint(points+i);
  }
2211 2212

  taosArrayDestroy(lpPoints);
S
shenglian zhou 已提交
2213

2214
  tfree(info);
2215
  return code;
2216 2217
}

2218
/**
2219
 * taos_schemaless_insert() parse and insert data points into database according to
2220 2221 2222 2223 2224 2225 2226 2227 2228 2229 2230 2231 2232 2233 2234 2235 2236 2237 2238
 * different protocol.
 *
 * @param $lines input array may contain multiple lines, each line indicates a data point.
 *               If protocol=2 is used input array should contain single JSON
 *               string(e.g. char *lines[] = {"$JSON_string"}). If need to insert
 *               multiple data points in JSON format, should include them in $JSON_string
 *               as a JSON array.
 * @param $numLines indicates how many data points in $lines.
 *                  If protocol = 2 is used this param will be ignored as $lines should
 *                  contain single JSON string.
 * @param $protocol indicates which protocol to use for parsing:
 *                  0 - influxDB line protocol
 *                  1 - OpenTSDB telnet line protocol
 *                  2 - OpenTSDB JSON format protocol
 * @return return zero for successful insertion. Otherwise return none-zero error code of
 *         failure reason.
 *
 */

2239
int taos_schemaless_insert(TAOS* taos, char* lines[], int numLines, int protocol) {
2240 2241 2242
  int code;
  switch (protocol) {
    case SML_LINE_PROTOCOL:
2243
      code = taos_insert_lines(taos, lines, numLines);
2244 2245 2246 2247 2248 2249 2250 2251 2252 2253 2254 2255 2256 2257
      break;
    case SML_TELNET_PROTOCOL:
      code = taos_insert_telnet_lines(taos, lines, numLines);
      break;
    case SML_JSON_PROTOCOL:
      code = taos_insert_json_payload(taos, *lines);
      break;
    default:
      code = TSDB_CODE_TSC_INVALID_PROTOCOL_TYPE;
      break;
  }

  return code;
}