sclfunc.c 90.2 KB
Newer Older
H
Hongze Cheng 已提交
1
#include "cJSON.h"
H
Haojun Liao 已提交
2 3
#include "function.h"
#include "scalar.h"
H
Haojun Liao 已提交
4
#include "sclInt.h"
D
dapan1121 已提交
5
#include "sclvector.h"
6
#include "tdatablock.h"
wmmhello's avatar
wmmhello 已提交
7
#include "tjson.h"
8
#include "ttime.h"
9

G
Ganlin Zhao 已提交
10 11 12 13
typedef float (*_float_fn)(float);
typedef double (*_double_fn)(double);
typedef double (*_double_fn_2)(double, double);
typedef int (*_conv_fn)(int);
H
Hongze Cheng 已提交
14
typedef void (*_trim_fn)(char *, char *, int32_t, int32_t);
G
Ganlin Zhao 已提交
15
typedef int16_t (*_len_fn)(char *, int32_t);
G
Ganlin Zhao 已提交
16

17
/** Math functions **/
H
Hongze Cheng 已提交
18
static double tlog(double v) { return log(v); }
19 20

static double tlog2(double v, double base) {
21 22 23 24 25 26
  double a = log(v);
  double b = log(base);
  if (isnan(a) || isinf(a)) {
    return a;
  } else if (isnan(b) || isinf(b)) {
    return b;
G
Ganlin Zhao 已提交
27 28
  } else if (b == 0) {
    return INFINITY;
29 30
  } else {
    return a / b;
31
  }
32 33
}

34
int32_t absFunction(SScalarParam *pInput, int32_t inputNum, SScalarParam *pOutput) {
H
Hongze Cheng 已提交
35
  SColumnInfoData *pInputData = pInput->columnData;
H
Haojun Liao 已提交
36 37 38
  SColumnInfoData *pOutputData = pOutput->columnData;

  int32_t type = GET_PARAM_TYPE(pInput);
39

H
Haojun Liao 已提交
40
  switch (type) {
41
    case TSDB_DATA_TYPE_FLOAT: {
H
Hongze Cheng 已提交
42
      float *in = (float *)pInputData->pData;
H
Haojun Liao 已提交
43
      float *out = (float *)pOutputData->pData;
44
      for (int32_t i = 0; i < pInput->numOfRows; ++i) {
45 46
        if (colDataIsNull_s(pInputData, i)) {
          colDataAppendNULL(pOutputData, i);
H
Haojun Liao 已提交
47 48
          continue;
        }
H
Hongze Cheng 已提交
49
        out[i] = (in[i] >= 0) ? in[i] : -in[i];
50
      }
H
Haojun Liao 已提交
51
      break;
52 53 54
    }

    case TSDB_DATA_TYPE_DOUBLE: {
H
Hongze Cheng 已提交
55
      double *in = (double *)pInputData->pData;
H
Haojun Liao 已提交
56
      double *out = (double *)pOutputData->pData;
57
      for (int32_t i = 0; i < pInput->numOfRows; ++i) {
58 59
        if (colDataIsNull_s(pInputData, i)) {
          colDataAppendNULL(pOutputData, i);
H
Haojun Liao 已提交
60 61
          continue;
        }
H
Hongze Cheng 已提交
62
        out[i] = (in[i] >= 0) ? in[i] : -in[i];
H
Haojun Liao 已提交
63 64
      }
      break;
65 66
    }

H
Haojun Liao 已提交
67
    case TSDB_DATA_TYPE_TINYINT: {
H
Hongze Cheng 已提交
68
      int8_t *in = (int8_t *)pInputData->pData;
H
Haojun Liao 已提交
69
      int8_t *out = (int8_t *)pOutputData->pData;
70
      for (int32_t i = 0; i < pInput->numOfRows; ++i) {
71 72
        if (colDataIsNull_s(pInputData, i)) {
          colDataAppendNULL(pOutputData, i);
H
Haojun Liao 已提交
73 74
          continue;
        }
H
Hongze Cheng 已提交
75
        out[i] = (in[i] >= 0) ? in[i] : -in[i];
76
      }
H
Haojun Liao 已提交
77 78
      break;
    }
79

H
Haojun Liao 已提交
80
    case TSDB_DATA_TYPE_SMALLINT: {
H
Hongze Cheng 已提交
81
      int16_t *in = (int16_t *)pInputData->pData;
H
Haojun Liao 已提交
82
      int16_t *out = (int16_t *)pOutputData->pData;
83
      for (int32_t i = 0; i < pInput->numOfRows; ++i) {
84 85
        if (colDataIsNull_s(pInputData, i)) {
          colDataAppendNULL(pOutputData, i);
H
Haojun Liao 已提交
86 87
          continue;
        }
H
Hongze Cheng 已提交
88
        out[i] = (in[i] >= 0) ? in[i] : -in[i];
89
      }
H
Haojun Liao 已提交
90 91
      break;
    }
92

H
Haojun Liao 已提交
93
    case TSDB_DATA_TYPE_INT: {
H
Hongze Cheng 已提交
94
      int32_t *in = (int32_t *)pInputData->pData;
H
Haojun Liao 已提交
95
      int32_t *out = (int32_t *)pOutputData->pData;
96
      for (int32_t i = 0; i < pInput->numOfRows; ++i) {
97 98
        if (colDataIsNull_s(pInputData, i)) {
          colDataAppendNULL(pOutputData, i);
H
Haojun Liao 已提交
99 100
          continue;
        }
H
Hongze Cheng 已提交
101
        out[i] = (in[i] >= 0) ? in[i] : -in[i];
102
      }
H
Haojun Liao 已提交
103 104
      break;
    }
105

H
Haojun Liao 已提交
106
    case TSDB_DATA_TYPE_BIGINT: {
H
Hongze Cheng 已提交
107
      int64_t *in = (int64_t *)pInputData->pData;
H
Haojun Liao 已提交
108
      int64_t *out = (int64_t *)pOutputData->pData;
109
      for (int32_t i = 0; i < pInput->numOfRows; ++i) {
110 111
        if (colDataIsNull_s(pInputData, i)) {
          colDataAppendNULL(pOutputData, i);
H
Haojun Liao 已提交
112 113
          continue;
        }
H
Hongze Cheng 已提交
114
        out[i] = (in[i] >= 0) ? in[i] : -in[i];
115
      }
H
Haojun Liao 已提交
116
      break;
117 118
    }

119 120 121 122 123 124 125
    case TSDB_DATA_TYPE_NULL: {
      for (int32_t i = 0; i < pInput->numOfRows; ++i) {
        colDataAppendNULL(pOutputData, i);
      }
      break;
    }

H
Haojun Liao 已提交
126
    default: {
127
      colDataAssign(pOutputData, pInputData, pInput->numOfRows, NULL);
128 129
    }
  }
130

H
Haojun Liao 已提交
131
  pOutput->numOfRows = pInput->numOfRows;
132
  return TSDB_CODE_SUCCESS;
133 134
}

H
Hongze Cheng 已提交
135
static int32_t doScalarFunctionUnique(SScalarParam *pInput, int32_t inputNum, SScalarParam *pOutput, _double_fn valFn) {
H
Haojun Liao 已提交
136
  int32_t type = GET_PARAM_TYPE(pInput);
137

138
  SColumnInfoData *pInputData = pInput->columnData;
H
Haojun Liao 已提交
139
  SColumnInfoData *pOutputData = pOutput->columnData;
140

141
  _getDoubleValue_fn_t getValueFn = getVectorDoubleValueFn(type);
142

143
  double *out = (double *)pOutputData->pData;
144

145
  for (int32_t i = 0; i < pInput->numOfRows; ++i) {
146
    if (colDataIsNull_s(pInputData, i) || IS_NULL_TYPE(type)) {
147
      colDataAppendNULL(pOutputData, i);
148
      continue;
149
    }
150 151 152 153 154 155
    double result = valFn(getValueFn(pInputData->pData, i));
    if (isinf(result) || isnan(result)) {
      colDataAppendNULL(pOutputData, i);
    } else {
      out[i] = result;
    }
156
  }
157

H
Haojun Liao 已提交
158
  pOutput->numOfRows = pInput->numOfRows;
159
  return TSDB_CODE_SUCCESS;
160 161
}

H
Hongze Cheng 已提交
162 163 164 165
static int32_t doScalarFunctionUnique2(SScalarParam *pInput, int32_t inputNum, SScalarParam *pOutput,
                                       _double_fn_2 valFn) {
  SColumnInfoData     *pInputData[2];
  SColumnInfoData     *pOutputData = pOutput->columnData;
G
Ganlin Zhao 已提交
166
  _getDoubleValue_fn_t getValueFn[2];
167

G
Ganlin Zhao 已提交
168 169
  for (int32_t i = 0; i < inputNum; ++i) {
    pInputData[i] = pInput[i].columnData;
H
Hongze Cheng 已提交
170
    getValueFn[i] = getVectorDoubleValueFn(GET_PARAM_TYPE(&pInput[i]));
G
Ganlin Zhao 已提交
171
  }
172

173
  double *out = (double *)pOutputData->pData;
H
Hongze Cheng 已提交
174
  double  result;
175

H
Hongze Cheng 已提交
176
  bool hasNullType = (IS_NULL_TYPE(GET_PARAM_TYPE(&pInput[0])) || IS_NULL_TYPE(GET_PARAM_TYPE(&pInput[1])));
177

wafwerar's avatar
wafwerar 已提交
178
  int32_t numOfRows = TMAX(pInput[0].numOfRows, pInput[1].numOfRows);
179 180
  if (pInput[0].numOfRows == pInput[1].numOfRows) {
    for (int32_t i = 0; i < numOfRows; ++i) {
H
Hongze Cheng 已提交
181
      if (colDataIsNull_s(pInputData[0], i) || colDataIsNull_s(pInputData[1], i) || hasNullType) {
182 183 184 185 186 187 188 189 190
        colDataAppendNULL(pOutputData, i);
        continue;
      }
      result = valFn(getValueFn[0](pInputData[0]->pData, i), getValueFn[1](pInputData[1]->pData, i));
      if (isinf(result) || isnan(result)) {
        colDataAppendNULL(pOutputData, i);
      } else {
        out[i] = result;
      }
191
    }
H
Hongze Cheng 已提交
192
  } else if (pInput[0].numOfRows == 1) {  // left operand is constant
193
    if (colDataIsNull_s(pInputData[0], 0) || hasNullType) {
194
      colDataAppendNNULL(pOutputData, 0, pInput[1].numOfRows);
195
    } else {
196 197 198 199 200
      for (int32_t i = 0; i < numOfRows; ++i) {
        if (colDataIsNull_s(pInputData[1], i)) {
          colDataAppendNULL(pOutputData, i);
          continue;
        }
201

202 203 204 205 206 207 208 209
        result = valFn(getValueFn[0](pInputData[0]->pData, 0), getValueFn[1](pInputData[1]->pData, i));
        if (isinf(result) || isnan(result)) {
          colDataAppendNULL(pOutputData, i);
          continue;
        }

        out[i] = result;
      }
210
    }
211
  } else if (pInput[1].numOfRows == 1) {
212
    if (colDataIsNull_s(pInputData[1], 0) || hasNullType) {
213
      colDataAppendNNULL(pOutputData, 0, pInput[0].numOfRows);
214
    } else {
215 216 217 218 219 220 221 222 223 224 225 226 227 228
      for (int32_t i = 0; i < numOfRows; ++i) {
        if (colDataIsNull_s(pInputData[0], i)) {
          colDataAppendNULL(pOutputData, i);
          continue;
        }

        result = valFn(getValueFn[0](pInputData[0]->pData, i), getValueFn[1](pInputData[1]->pData, 0));
        if (isinf(result) || isnan(result)) {
          colDataAppendNULL(pOutputData, i);
          continue;
        }

        out[i] = result;
      }
229
    }
230
  }
231

232
  pOutput->numOfRows = numOfRows;
233 234 235
  return TSDB_CODE_SUCCESS;
}

H
Hongze Cheng 已提交
236 237
static int32_t doScalarFunction(SScalarParam *pInput, int32_t inputNum, SScalarParam *pOutput, _float_fn f1,
                                _double_fn d1) {
H
Haojun Liao 已提交
238
  int32_t type = GET_PARAM_TYPE(pInput);
239

H
Hongze Cheng 已提交
240
  SColumnInfoData *pInputData = pInput->columnData;
H
Haojun Liao 已提交
241
  SColumnInfoData *pOutputData = pOutput->columnData;
242

H
Haojun Liao 已提交
243 244
  switch (type) {
    case TSDB_DATA_TYPE_FLOAT: {
H
Hongze Cheng 已提交
245
      float *in = (float *)pInputData->pData;
H
Haojun Liao 已提交
246
      float *out = (float *)pOutputData->pData;
247

248
      for (int32_t i = 0; i < pInput->numOfRows; ++i) {
249 250
        if (colDataIsNull_s(pInputData, i)) {
          colDataAppendNULL(pOutputData, i);
H
Haojun Liao 已提交
251 252
          continue;
        }
253
        out[i] = f1(in[i]);
254
      }
H
Haojun Liao 已提交
255 256
      break;
    }
257

H
Haojun Liao 已提交
258
    case TSDB_DATA_TYPE_DOUBLE: {
H
Hongze Cheng 已提交
259
      double *in = (double *)pInputData->pData;
H
Haojun Liao 已提交
260
      double *out = (double *)pOutputData->pData;
261

262
      for (int32_t i = 0; i < pInput->numOfRows; ++i) {
263 264
        if (colDataIsNull_s(pInputData, i)) {
          colDataAppendNULL(pOutputData, i);
H
Haojun Liao 已提交
265 266
          continue;
        }
267
        out[i] = d1(in[i]);
268
      }
H
Haojun Liao 已提交
269 270 271
      break;
    }

272 273 274 275 276 277 278
    case TSDB_DATA_TYPE_NULL: {
      for (int32_t i = 0; i < pInput->numOfRows; ++i) {
        colDataAppendNULL(pOutputData, i);
      }
      break;
    }

H
Haojun Liao 已提交
279
    default: {
280
      colDataAssign(pOutputData, pInputData, pInput->numOfRows, NULL);
281
    }
282 283
  }

H
Haojun Liao 已提交
284
  pOutput->numOfRows = pInput->numOfRows;
285 286
  return TSDB_CODE_SUCCESS;
}
287

G
Ganlin Zhao 已提交
288
/** String functions **/
H
Hongze Cheng 已提交
289
static int16_t tlength(char *input, int32_t type) { return varDataLen(input); }
G
Ganlin Zhao 已提交
290

291
static int16_t tcharlength(char *input, int32_t type) {
G
Ganlin Zhao 已提交
292 293
  if (type == TSDB_DATA_TYPE_VARCHAR) {
    return varDataLen(input);
H
Hongze Cheng 已提交
294
  } else {  // NCHAR
G
Ganlin Zhao 已提交
295
    return varDataLen(input) / TSDB_NCHAR_SIZE;
G
Ganlin Zhao 已提交
296
  }
G
Ganlin Zhao 已提交
297
}
G
Ganlin Zhao 已提交
298

299
static void tltrim(char *input, char *output, int32_t type, int32_t charLen) {
G
Ganlin Zhao 已提交
300 301 302 303 304 305 306 307
  int32_t numOfSpaces = 0;
  if (type == TSDB_DATA_TYPE_VARCHAR) {
    for (int32_t i = 0; i < charLen; ++i) {
      if (!isspace(*(varDataVal(input) + i))) {
        break;
      }
      numOfSpaces++;
    }
H
Hongze Cheng 已提交
308
  } else {  // NCHAR
G
Ganlin Zhao 已提交
309 310 311 312 313 314 315
    for (int32_t i = 0; i < charLen; ++i) {
      if (!iswspace(*((uint32_t *)varDataVal(input) + i))) {
        break;
      }
      numOfSpaces++;
    }
  }
G
Ganlin Zhao 已提交
316

G
Ganlin Zhao 已提交
317 318 319 320 321 322 323 324
  int32_t resLen;
  if (type == TSDB_DATA_TYPE_VARCHAR) {
    resLen = charLen - numOfSpaces;
    memcpy(varDataVal(output), varDataVal(input) + numOfSpaces, resLen);
  } else {
    resLen = (charLen - numOfSpaces) * TSDB_NCHAR_SIZE;
    memcpy(varDataVal(output), varDataVal(input) + numOfSpaces * TSDB_NCHAR_SIZE, resLen);
  }
G
Ganlin Zhao 已提交
325

G
Ganlin Zhao 已提交
326 327 328
  varDataSetLen(output, resLen);
}

329
static void trtrim(char *input, char *output, int32_t type, int32_t charLen) {
G
Ganlin Zhao 已提交
330 331 332 333 334 335 336 337
  int32_t numOfSpaces = 0;
  if (type == TSDB_DATA_TYPE_VARCHAR) {
    for (int32_t i = charLen - 1; i >= 0; --i) {
      if (!isspace(*(varDataVal(input) + i))) {
        break;
      }
      numOfSpaces++;
    }
H
Hongze Cheng 已提交
338
  } else {  // NCHAR
339
    for (int32_t i = charLen - 1; i >= 0; --i) {
G
Ganlin Zhao 已提交
340 341 342 343
      if (!iswspace(*((uint32_t *)varDataVal(input) + i))) {
        break;
      }
      numOfSpaces++;
G
Ganlin Zhao 已提交
344
    }
G
Ganlin Zhao 已提交
345
  }
G
Ganlin Zhao 已提交
346

G
Ganlin Zhao 已提交
347 348 349 350 351
  int32_t resLen;
  if (type == TSDB_DATA_TYPE_VARCHAR) {
    resLen = charLen - numOfSpaces;
  } else {
    resLen = (charLen - numOfSpaces) * TSDB_NCHAR_SIZE;
G
Ganlin Zhao 已提交
352
  }
G
Ganlin Zhao 已提交
353
  memcpy(varDataVal(output), varDataVal(input), resLen);
G
Ganlin Zhao 已提交
354

G
Ganlin Zhao 已提交
355
  varDataSetLen(output, resLen);
G
Ganlin Zhao 已提交
356 357
}

358
static int32_t doLengthFunction(SScalarParam *pInput, int32_t inputNum, SScalarParam *pOutput, _len_fn lenFn) {
G
Ganlin Zhao 已提交
359 360
  int32_t type = GET_PARAM_TYPE(pInput);

H
Hongze Cheng 已提交
361
  SColumnInfoData *pInputData = pInput->columnData;
G
Ganlin Zhao 已提交
362 363
  SColumnInfoData *pOutputData = pOutput->columnData;

364 365
  ASSERT(pOutputData->info.type == TSDB_DATA_TYPE_BIGINT);
  int64_t *out = (int64_t *)pOutputData->pData;
G
Ganlin Zhao 已提交
366 367

  for (int32_t i = 0; i < pInput->numOfRows; ++i) {
368
    if (colDataIsNull_s(pInputData, i)) {
369
      colDataAppendNULL(pOutputData, i);
G
Ganlin Zhao 已提交
370 371 372
      continue;
    }

373
    char *in = colDataGetData(pInputData, i);
G
Ganlin Zhao 已提交
374
    out[i] = lenFn(in, type);
G
Ganlin Zhao 已提交
375 376 377 378 379
  }

  pOutput->numOfRows = pInput->numOfRows;
  return TSDB_CODE_SUCCESS;
}
380

381 382
static int32_t concatCopyHelper(const char *input, char *output, bool hasNchar, int32_t type, int16_t *dataLen) {
  if (hasNchar && type == TSDB_DATA_TYPE_VARCHAR) {
383
    TdUcs4 *newBuf = taosMemoryCalloc((varDataLen(input) + 1) * TSDB_NCHAR_SIZE, 1);
384
    int32_t len = varDataLen(input);
H
Hongze Cheng 已提交
385
    bool    ret = taosMbsToUcs4(varDataVal(input), len, newBuf, (varDataLen(input) + 1) * TSDB_NCHAR_SIZE, &len);
386 387 388 389 390 391 392 393 394 395 396 397
    if (!ret) {
      taosMemoryFree(newBuf);
      return TSDB_CODE_FAILED;
    }
    memcpy(varDataVal(output) + *dataLen, newBuf, varDataLen(input) * TSDB_NCHAR_SIZE);
    *dataLen += varDataLen(input) * TSDB_NCHAR_SIZE;
    taosMemoryFree(newBuf);
  } else {
    memcpy(varDataVal(output) + *dataLen, varDataVal(input), varDataLen(input));
    *dataLen += varDataLen(input);
  }
  return TSDB_CODE_SUCCESS;
398
}
G
Ganlin Zhao 已提交
399

400 401 402 403 404 405 406 407 408 409 410 411 412
static int32_t getNumOfNullEntries(SColumnInfoData *pColumnInfoData, int32_t numOfRows) {
  int32_t numOfNulls = 0;
  if (!pColumnInfoData->hasNull) {
    return numOfNulls;
  }
  for (int i = 0; i < numOfRows; ++i) {
    if (pColumnInfoData->varmeta.offset[i] == -1) {
      numOfNulls++;
    }
  }
  return numOfNulls;
}

G
Ganlin Zhao 已提交
413
int32_t concatFunction(SScalarParam *pInput, int32_t inputNum, SScalarParam *pOutput) {
H
Hongze Cheng 已提交
414
  int32_t           ret = TSDB_CODE_SUCCESS;
G
Ganlin Zhao 已提交
415
  SColumnInfoData **pInputData = taosMemoryCalloc(inputNum, sizeof(SColumnInfoData *));
H
Hongze Cheng 已提交
416 417 418
  SColumnInfoData  *pOutputData = pOutput->columnData;
  char            **input = taosMemoryCalloc(inputNum, POINTER_BYTES);
  char             *outputBuf = NULL;
G
Ganlin Zhao 已提交
419

420
  int32_t inputLen = 0;
421
  int32_t numOfRows = 0;
H
Hongze Cheng 已提交
422
  bool    hasNchar = (GET_PARAM_TYPE(pOutput) == TSDB_DATA_TYPE_NCHAR) ? true : false;
423 424 425 426 427
  for (int32_t i = 0; i < inputNum; ++i) {
    if (pInput[i].numOfRows > numOfRows) {
      numOfRows = pInput[i].numOfRows;
    }
  }
G
Ganlin Zhao 已提交
428 429
  for (int32_t i = 0; i < inputNum; ++i) {
    pInputData[i] = pInput[i].columnData;
430
    int32_t factor = 1;
431
    if (hasNchar && (GET_PARAM_TYPE(&pInput[i]) == TSDB_DATA_TYPE_VARCHAR)) {
432 433
      factor = TSDB_NCHAR_SIZE;
    }
434 435

    int32_t numOfNulls = getNumOfNullEntries(pInputData[i], pInput[i].numOfRows);
436
    if (pInput[i].numOfRows == 1) {
437
      inputLen += (pInputData[i]->varmeta.length - VARSTR_HEADER_SIZE) * factor * (numOfRows - numOfNulls);
438
    } else {
439
      inputLen += (pInputData[i]->varmeta.length - (numOfRows - numOfNulls) * VARSTR_HEADER_SIZE) * factor;
440
    }
441 442
  }

443 444 445
  int32_t outputLen = inputLen + numOfRows * VARSTR_HEADER_SIZE;
  outputBuf = taosMemoryCalloc(outputLen, 1);
  char *output = outputBuf;
G
Ganlin Zhao 已提交
446

447
  for (int32_t k = 0; k < numOfRows; ++k) {
448
    bool hasNull = false;
G
Ganlin Zhao 已提交
449
    for (int32_t i = 0; i < inputNum; ++i) {
H
Hongze Cheng 已提交
450
      if (colDataIsNull_s(pInputData[i], k) || IS_NULL_TYPE(GET_PARAM_TYPE(&pInput[i]))) {
451
        colDataAppendNULL(pOutputData, k);
G
Ganlin Zhao 已提交
452 453 454 455 456 457 458 459 460 461 462
        hasNull = true;
        break;
      }
    }

    if (hasNull) {
      continue;
    }

    int16_t dataLen = 0;
    for (int32_t i = 0; i < inputNum; ++i) {
463 464
      int32_t rowIdx = (pInput[i].numOfRows == 1) ? 0 : k;
      input[i] = colDataGetData(pInputData[i], rowIdx);
465 466 467 468

      ret = concatCopyHelper(input[i], output, hasNchar, GET_PARAM_TYPE(&pInput[i]), &dataLen);
      if (ret != TSDB_CODE_SUCCESS) {
        goto DONE;
469
      }
G
Ganlin Zhao 已提交
470
    }
471
    varDataSetLen(output, dataLen);
472 473
    colDataAppend(pOutputData, k, output, false);
    output += varDataTLen(output);
G
Ganlin Zhao 已提交
474 475
  }

476
  pOutput->numOfRows = numOfRows;
477 478

DONE:
479
  taosMemoryFree(input);
480
  taosMemoryFree(outputBuf);
G
Ganlin Zhao 已提交
481 482
  taosMemoryFree(pInputData);

483
  return ret;
G
Ganlin Zhao 已提交
484 485 486
}

int32_t concatWsFunction(SScalarParam *pInput, int32_t inputNum, SScalarParam *pOutput) {
H
Hongze Cheng 已提交
487
  int32_t           ret = TSDB_CODE_SUCCESS;
G
Ganlin Zhao 已提交
488
  SColumnInfoData **pInputData = taosMemoryCalloc(inputNum, sizeof(SColumnInfoData *));
H
Hongze Cheng 已提交
489 490 491
  SColumnInfoData  *pOutputData = pOutput->columnData;
  char            **input = taosMemoryCalloc(inputNum, POINTER_BYTES);
  char             *outputBuf = NULL;
G
Ganlin Zhao 已提交
492

493
  int32_t inputLen = 0;
494
  int32_t numOfRows = 0;
H
Hongze Cheng 已提交
495
  bool    hasNchar = (GET_PARAM_TYPE(pOutput) == TSDB_DATA_TYPE_NCHAR) ? true : false;
496
  for (int32_t i = 1; i < inputNum; ++i) {
497 498 499 500
    if (pInput[i].numOfRows > numOfRows) {
      numOfRows = pInput[i].numOfRows;
    }
  }
G
Ganlin Zhao 已提交
501 502
  for (int32_t i = 0; i < inputNum; ++i) {
    pInputData[i] = pInput[i].columnData;
503
    int32_t factor = 1;
504
    if (hasNchar && (GET_PARAM_TYPE(&pInput[i]) == TSDB_DATA_TYPE_VARCHAR)) {
505 506
      factor = TSDB_NCHAR_SIZE;
    }
507 508

    int32_t numOfNulls = getNumOfNullEntries(pInputData[i], pInput[i].numOfRows);
509 510
    if (i == 0) {
      // calculate required separator space
H
Hongze Cheng 已提交
511 512
      inputLen +=
          (pInputData[0]->varmeta.length - VARSTR_HEADER_SIZE) * (numOfRows - numOfNulls) * (inputNum - 2) * factor;
513
    } else if (pInput[i].numOfRows == 1) {
514
      inputLen += (pInputData[i]->varmeta.length - VARSTR_HEADER_SIZE) * (numOfRows - numOfNulls) * factor;
515
    } else {
516
      inputLen += (pInputData[i]->varmeta.length - (numOfRows - numOfNulls) * VARSTR_HEADER_SIZE) * factor;
517 518 519
    }
  }

520 521 522
  int32_t outputLen = inputLen + numOfRows * VARSTR_HEADER_SIZE;
  outputBuf = taosMemoryCalloc(outputLen, 1);
  char *output = outputBuf;
G
Ganlin Zhao 已提交
523

524
  for (int32_t k = 0; k < numOfRows; ++k) {
H
Hongze Cheng 已提交
525
    if (colDataIsNull_s(pInputData[0], k) || IS_NULL_TYPE(GET_PARAM_TYPE(&pInput[0]))) {
526
      colDataAppendNULL(pOutputData, k);
G
Ganlin Zhao 已提交
527 528 529 530
      continue;
    }

    int16_t dataLen = 0;
H
Hongze Cheng 已提交
531
    bool    hasNull = false;
G
Ganlin Zhao 已提交
532
    for (int32_t i = 1; i < inputNum; ++i) {
H
Hongze Cheng 已提交
533
      if (colDataIsNull_s(pInputData[i], k) || IS_NULL_TYPE(GET_PARAM_TYPE(&pInput[i]))) {
534 535
        hasNull = true;
        break;
G
Ganlin Zhao 已提交
536 537
      }

538 539
      int32_t rowIdx = (pInput[i].numOfRows == 1) ? 0 : k;
      input[i] = colDataGetData(pInputData[i], rowIdx);
540

541 542 543
      ret = concatCopyHelper(input[i], output, hasNchar, GET_PARAM_TYPE(&pInput[i]), &dataLen);
      if (ret != TSDB_CODE_SUCCESS) {
        goto DONE;
544
      }
G
Ganlin Zhao 已提交
545 546

      if (i < inputNum - 1) {
H
Hongze Cheng 已提交
547
        // insert the separator
548
        char *sep = (pInput[0].numOfRows == 1) ? colDataGetData(pInputData[0], 0) : colDataGetData(pInputData[0], k);
H
Hongze Cheng 已提交
549
        ret = concatCopyHelper(sep, output, hasNchar, GET_PARAM_TYPE(&pInput[0]), &dataLen);
550
        if (ret != TSDB_CODE_SUCCESS) {
551
          goto DONE;
552
        }
G
Ganlin Zhao 已提交
553 554
      }
    }
555 556 557 558 559 560 561 562 563

    if (hasNull) {
      colDataAppendNULL(pOutputData, k);
      memset(output, 0, dataLen);
    } else {
      varDataSetLen(output, dataLen);
      colDataAppend(pOutputData, k, output, false);
      output += varDataTLen(output);
    }
G
Ganlin Zhao 已提交
564 565
  }

566
  pOutput->numOfRows = numOfRows;
567 568

DONE:
569
  taosMemoryFree(input);
570
  taosMemoryFree(outputBuf);
G
Ganlin Zhao 已提交
571 572
  taosMemoryFree(pInputData);

573
  return ret;
G
Ganlin Zhao 已提交
574 575
}

576
static int32_t doCaseConvFunction(SScalarParam *pInput, int32_t inputNum, SScalarParam *pOutput, _conv_fn convFn) {
G
Ganlin Zhao 已提交
577 578
  int32_t type = GET_PARAM_TYPE(pInput);

H
Hongze Cheng 已提交
579
  SColumnInfoData *pInputData = pInput->columnData;
G
Ganlin Zhao 已提交
580 581
  SColumnInfoData *pOutputData = pOutput->columnData;

582
  int32_t outputLen = pInputData->varmeta.length;
H
Hongze Cheng 已提交
583 584
  char   *outputBuf = taosMemoryCalloc(outputLen, 1);
  char   *output = outputBuf;
585

G
Ganlin Zhao 已提交
586
  for (int32_t i = 0; i < pInput->numOfRows; ++i) {
587 588
    if (colDataIsNull_s(pInputData, i)) {
      colDataAppendNULL(pOutputData, i);
G
Ganlin Zhao 已提交
589 590 591
      continue;
    }

H
Hongze Cheng 已提交
592
    char   *input = colDataGetData(pInput[0].columnData, i);
593
    int32_t len = varDataLen(input);
G
Ganlin Zhao 已提交
594 595
    if (type == TSDB_DATA_TYPE_VARCHAR) {
      for (int32_t j = 0; j < len; ++j) {
596
        *(varDataVal(output) + j) = convFn(*(varDataVal(input) + j));
G
Ganlin Zhao 已提交
597
      }
H
Hongze Cheng 已提交
598
    } else {  // NCHAR
G
Ganlin Zhao 已提交
599
      for (int32_t j = 0; j < len / TSDB_NCHAR_SIZE; ++j) {
600
        *((uint32_t *)varDataVal(output) + j) = convFn(*((uint32_t *)varDataVal(input) + j));
G
Ganlin Zhao 已提交
601 602
      }
    }
603
    varDataSetLen(output, len);
604
    colDataAppend(pOutputData, i, output, false);
605
    output += varDataTLen(output);
G
Ganlin Zhao 已提交
606 607 608
  }

  pOutput->numOfRows = pInput->numOfRows;
609
  taosMemoryFree(outputBuf);
G
Ganlin Zhao 已提交
610 611 612 613

  return TSDB_CODE_SUCCESS;
}

614
static int32_t doTrimFunction(SScalarParam *pInput, int32_t inputNum, SScalarParam *pOutput, _trim_fn trimFn) {
G
Ganlin Zhao 已提交
615 616
  int32_t type = GET_PARAM_TYPE(pInput);

H
Hongze Cheng 已提交
617
  SColumnInfoData *pInputData = pInput->columnData;
G
Ganlin Zhao 已提交
618 619
  SColumnInfoData *pOutputData = pOutput->columnData;

620
  int32_t outputLen = pInputData->varmeta.length;
H
Hongze Cheng 已提交
621 622
  char   *outputBuf = taosMemoryCalloc(outputLen, 1);
  char   *output = outputBuf;
623

G
Ganlin Zhao 已提交
624
  for (int32_t i = 0; i < pInput->numOfRows; ++i) {
625 626
    if (colDataIsNull_s(pInputData, i)) {
      colDataAppendNULL(pOutputData, i);
G
Ganlin Zhao 已提交
627 628 629
      continue;
    }

H
Hongze Cheng 已提交
630
    char   *input = colDataGetData(pInputData, i);
631
    int32_t len = varDataLen(input);
G
Ganlin Zhao 已提交
632
    int32_t charLen = (type == TSDB_DATA_TYPE_VARCHAR) ? len : len / TSDB_NCHAR_SIZE;
633 634
    trimFn(input, output, type, charLen);

635
    colDataAppend(pOutputData, i, output, false);
636
    output += varDataTLen(output);
G
Ganlin Zhao 已提交
637 638 639
  }

  pOutput->numOfRows = pInput->numOfRows;
640
  taosMemoryFree(outputBuf);
G
Ganlin Zhao 已提交
641 642 643 644 645 646 647 648 649 650 651 652 653 654

  return TSDB_CODE_SUCCESS;
}

int32_t substrFunction(SScalarParam *pInput, int32_t inputNum, SScalarParam *pOutput) {
  int32_t subPos = 0;
  GET_TYPED_DATA(subPos, int32_t, GET_PARAM_TYPE(&pInput[1]), pInput[1].columnData->pData);

  int32_t subLen = INT16_MAX;
  if (inputNum == 3) {
    GET_TYPED_DATA(subLen, int32_t, GET_PARAM_TYPE(&pInput[2]), pInput[2].columnData->pData);
    subLen = (GET_PARAM_TYPE(pInput) == TSDB_DATA_TYPE_VARCHAR) ? subLen : subLen * TSDB_NCHAR_SIZE;
  }

H
Hongze Cheng 已提交
655
  SColumnInfoData *pInputData = pInput->columnData;
G
Ganlin Zhao 已提交
656 657
  SColumnInfoData *pOutputData = pOutput->columnData;

658
  int32_t outputLen = pInputData->varmeta.length * pInput->numOfRows;
H
Hongze Cheng 已提交
659 660
  char   *outputBuf = taosMemoryCalloc(outputLen, 1);
  char   *output = outputBuf;
661 662

  for (int32_t i = 0; i < pInput->numOfRows; ++i) {
663 664
    if (colDataIsNull_s(pInputData, i)) {
      colDataAppendNULL(pOutputData, i);
G
Ganlin Zhao 已提交
665 666
      continue;
    }
H
Hongze Cheng 已提交
667
    char   *input = colDataGetData(pInput[0].columnData, i);
668
    int32_t len = varDataLen(input);
G
Ganlin Zhao 已提交
669 670 671 672
    int32_t startPosBytes;

    if (subPos > 0) {
      startPosBytes = (GET_PARAM_TYPE(pInput) == TSDB_DATA_TYPE_VARCHAR) ? subPos - 1 : (subPos - 1) * TSDB_NCHAR_SIZE;
wafwerar's avatar
wafwerar 已提交
673
      startPosBytes = TMIN(startPosBytes, len);
G
Ganlin Zhao 已提交
674
    } else {
H
Hongze Cheng 已提交
675 676
      startPosBytes =
          (GET_PARAM_TYPE(pInput) == TSDB_DATA_TYPE_VARCHAR) ? len + subPos : len + subPos * TSDB_NCHAR_SIZE;
wafwerar's avatar
wafwerar 已提交
677
      startPosBytes = TMAX(startPosBytes, 0);
G
Ganlin Zhao 已提交
678 679
    }

wafwerar's avatar
wafwerar 已提交
680
    int32_t resLen = TMIN(subLen, len - startPosBytes);
681 682
    if (resLen > 0) {
      memcpy(varDataVal(output), varDataVal(input) + startPosBytes, resLen);
G
Ganlin Zhao 已提交
683 684
    }

685
    varDataSetLen(output, resLen);
H
Hongze Cheng 已提交
686
    colDataAppend(pOutputData, i, output, false);
687
    output += varDataTLen(output);
G
Ganlin Zhao 已提交
688 689 690
  }

  pOutput->numOfRows = pInput->numOfRows;
691
  taosMemoryFree(outputBuf);
G
Ganlin Zhao 已提交
692 693 694 695

  return TSDB_CODE_SUCCESS;
}

696
/** Conversion functions **/
G
Ganlin Zhao 已提交
697
int32_t castFunction(SScalarParam *pInput, int32_t inputNum, SScalarParam *pOutput) {
H
Hongze Cheng 已提交
698 699
  int16_t inputType = GET_PARAM_TYPE(&pInput[0]);
  int16_t inputLen = GET_PARAM_BYTES(&pInput[0]);
700
  int16_t outputType = GET_PARAM_TYPE(&pOutput[0]);
H
Hongze Cheng 已提交
701
  int64_t outputLen = GET_PARAM_BYTES(&pOutput[0]);
G
Ganlin Zhao 已提交
702

703
  int32_t code = TSDB_CODE_SUCCESS;
H
Hongze Cheng 已提交
704 705
  char   *convBuf = taosMemoryMalloc(inputLen);
  char   *output = taosMemoryCalloc(1, outputLen + TSDB_NCHAR_SIZE);
706 707 708 709 710 711
  char    buf[400] = {0};

  if (convBuf == NULL || output == NULL) {
    code = TSDB_CODE_OUT_OF_MEMORY;
    goto _end;
  }
G
Ganlin Zhao 已提交
712 713 714 715 716 717

  for (int32_t i = 0; i < pInput[0].numOfRows; ++i) {
    if (colDataIsNull_s(pInput[0].columnData, i)) {
      colDataAppendNULL(pOutput->columnData, i);
      continue;
    }
718

719
    char *input = colDataGetData(pInput[0].columnData, i);
G
Ganlin Zhao 已提交
720

H
Hongze Cheng 已提交
721
    switch (outputType) {
722 723
      case TSDB_DATA_TYPE_TINYINT: {
        if (inputType == TSDB_DATA_TYPE_BINARY) {
724 725 726
          memcpy(buf, varDataVal(input), varDataLen(input));
          buf[varDataLen(input)] = 0;
          *(int8_t *)output = taosStr2Int8(buf, NULL, 10);
727
        } else if (inputType == TSDB_DATA_TYPE_NCHAR) {
H
Hongze Cheng 已提交
728
          int32_t len = taosUcs4ToMbs((TdUcs4 *)varDataVal(input), varDataLen(input), convBuf);
729
          if (len < 0) {
730 731
            code = TSDB_CODE_FAILED;
            goto _end;
732
          }
733 734 735

          convBuf[len] = 0;
          *(int8_t *)output = taosStr2Int8(convBuf, NULL, 10);
736 737 738 739 740 741 742
        } else {
          GET_TYPED_DATA(*(int8_t *)output, int8_t, inputType, input);
        }
        break;
      }
      case TSDB_DATA_TYPE_SMALLINT: {
        if (inputType == TSDB_DATA_TYPE_BINARY) {
743 744 745
          memcpy(buf, varDataVal(input), varDataLen(input));
          buf[varDataLen(input)] = 0;
          *(int16_t *)output = taosStr2Int16(buf, NULL, 10);
746
        } else if (inputType == TSDB_DATA_TYPE_NCHAR) {
H
Hongze Cheng 已提交
747
          int32_t len = taosUcs4ToMbs((TdUcs4 *)varDataVal(input), varDataLen(input), convBuf);
748
          if (len < 0) {
749 750
            code = TSDB_CODE_FAILED;
            goto _end;
751
          }
752 753
          convBuf[len] = 0;
          *(int16_t *)output = taosStr2Int16(convBuf, NULL, 10);
754 755 756 757 758 759 760
        } else {
          GET_TYPED_DATA(*(int16_t *)output, int16_t, inputType, input);
        }
        break;
      }
      case TSDB_DATA_TYPE_INT: {
        if (inputType == TSDB_DATA_TYPE_BINARY) {
761 762 763
          memcpy(buf, varDataVal(input), varDataLen(input));
          buf[varDataLen(input)] = 0;
          *(int32_t *)output = taosStr2Int32(buf, NULL, 10);
764
        } else if (inputType == TSDB_DATA_TYPE_NCHAR) {
H
Hongze Cheng 已提交
765
          int32_t len = taosUcs4ToMbs((TdUcs4 *)varDataVal(input), varDataLen(input), convBuf);
766
          if (len < 0) {
767 768
            code = TSDB_CODE_FAILED;
            goto _end;
769
          }
770 771 772

          convBuf[len] = 0;
          *(int32_t *)output = taosStr2Int32(convBuf, NULL, 10);
773 774 775 776 777
        } else {
          GET_TYPED_DATA(*(int32_t *)output, int32_t, inputType, input);
        }
        break;
      }
G
Ganlin Zhao 已提交
778 779
      case TSDB_DATA_TYPE_BIGINT: {
        if (inputType == TSDB_DATA_TYPE_BINARY) {
780 781 782
          memcpy(buf, varDataVal(input), varDataLen(input));
          buf[varDataLen(input)] = 0;
          *(int64_t *)output = taosStr2Int64(buf, NULL, 10);
G
Ganlin Zhao 已提交
783
        } else if (inputType == TSDB_DATA_TYPE_NCHAR) {
H
Hongze Cheng 已提交
784
          int32_t len = taosUcs4ToMbs((TdUcs4 *)varDataVal(input), varDataLen(input), convBuf);
G
Ganlin Zhao 已提交
785
          if (len < 0) {
786 787
            code = TSDB_CODE_FAILED;
            goto _end;
G
Ganlin Zhao 已提交
788
          }
789 790
          convBuf[len] = 0;
          *(int64_t *)output = taosStr2Int64(convBuf, NULL, 10);
G
Ganlin Zhao 已提交
791 792 793 794 795
        } else {
          GET_TYPED_DATA(*(int64_t *)output, int64_t, inputType, input);
        }
        break;
      }
796 797
      case TSDB_DATA_TYPE_UTINYINT: {
        if (inputType == TSDB_DATA_TYPE_BINARY) {
798 799 800
          memcpy(buf, varDataVal(input), varDataLen(input));
          buf[varDataLen(input)] = 0;
          *(uint8_t *)output = taosStr2UInt8(buf, NULL, 10);
801
        } else if (inputType == TSDB_DATA_TYPE_NCHAR) {
802
          int32_t len = taosUcs4ToMbs((TdUcs4 *)varDataVal(input), varDataLen(input), convBuf);
803
          if (len < 0) {
804 805
            code = TSDB_CODE_FAILED;
            goto _end;
806
          }
807 808
          convBuf[len] = 0;
          *(uint8_t *)output = taosStr2UInt8(convBuf, NULL, 10);
809 810 811 812 813 814 815
        } else {
          GET_TYPED_DATA(*(uint8_t *)output, uint8_t, inputType, input);
        }
        break;
      }
      case TSDB_DATA_TYPE_USMALLINT: {
        if (inputType == TSDB_DATA_TYPE_BINARY) {
816 817 818
          memcpy(buf, varDataVal(input), varDataLen(input));
          buf[varDataLen(input)] = 0;
          *(uint16_t *)output = taosStr2UInt16(buf, NULL, 10);
819
        } else if (inputType == TSDB_DATA_TYPE_NCHAR) {
820
          int32_t len = taosUcs4ToMbs((TdUcs4 *)varDataVal(input), varDataLen(input), convBuf);
821
          if (len < 0) {
822 823
            code = TSDB_CODE_FAILED;
            goto _end;
824
          }
825 826
          convBuf[len] = 0;
          *(uint16_t *)output = taosStr2UInt16(convBuf, NULL, 10);
827 828 829 830 831 832 833
        } else {
          GET_TYPED_DATA(*(uint16_t *)output, uint16_t, inputType, input);
        }
        break;
      }
      case TSDB_DATA_TYPE_UINT: {
        if (inputType == TSDB_DATA_TYPE_BINARY) {
834 835 836
          memcpy(buf, varDataVal(input), varDataLen(input));
          buf[varDataLen(input)] = 0;
          *(uint32_t *)output = taosStr2UInt32(buf, NULL, 10);
837
        } else if (inputType == TSDB_DATA_TYPE_NCHAR) {
838
          int32_t len = taosUcs4ToMbs((TdUcs4 *)varDataVal(input), varDataLen(input), convBuf);
839
          if (len < 0) {
840 841
            code = TSDB_CODE_FAILED;
            goto _end;
842
          }
843 844
          convBuf[len] = 0;
          *(uint32_t *)output = taosStr2UInt32(convBuf, NULL, 10);
845 846 847 848 849
        } else {
          GET_TYPED_DATA(*(uint32_t *)output, uint32_t, inputType, input);
        }
        break;
      }
G
Ganlin Zhao 已提交
850 851
      case TSDB_DATA_TYPE_UBIGINT: {
        if (inputType == TSDB_DATA_TYPE_BINARY) {
852 853 854
          memcpy(buf, varDataVal(input), varDataLen(input));
          buf[varDataLen(input)] = 0;
          *(uint64_t *)output = taosStr2UInt64(buf, NULL, 10);
G
Ganlin Zhao 已提交
855
        } else if (inputType == TSDB_DATA_TYPE_NCHAR) {
856
          int32_t len = taosUcs4ToMbs((TdUcs4 *)varDataVal(input), varDataLen(input), convBuf);
G
Ganlin Zhao 已提交
857
          if (len < 0) {
858 859
            code = TSDB_CODE_FAILED;
            goto _end;
G
Ganlin Zhao 已提交
860
          }
861 862 863

          convBuf[len] = 0;
          *(uint64_t *)output = taosStr2UInt64(convBuf, NULL, 10);
G
Ganlin Zhao 已提交
864 865 866 867 868
        } else {
          GET_TYPED_DATA(*(uint64_t *)output, uint64_t, inputType, input);
        }
        break;
      }
869 870
      case TSDB_DATA_TYPE_FLOAT: {
        if (inputType == TSDB_DATA_TYPE_BINARY) {
871 872 873
          memcpy(buf, varDataVal(input), varDataLen(input));
          buf[varDataLen(input)] = 0;
          *(float *)output = taosStr2Float(buf, NULL);
874
        } else if (inputType == TSDB_DATA_TYPE_NCHAR) {
H
Hongze Cheng 已提交
875
          int32_t len = taosUcs4ToMbs((TdUcs4 *)varDataVal(input), varDataLen(input), convBuf);
876
          if (len < 0) {
877 878
            code = TSDB_CODE_FAILED;
            goto _end;
879
          }
880 881
          convBuf[len] = 0;
          *(float *)output = taosStr2Float(convBuf, NULL);
882 883 884 885 886 887 888
        } else {
          GET_TYPED_DATA(*(float *)output, float, inputType, input);
        }
        break;
      }
      case TSDB_DATA_TYPE_DOUBLE: {
        if (inputType == TSDB_DATA_TYPE_BINARY) {
889 890 891
          memcpy(buf, varDataVal(input), varDataLen(input));
          buf[varDataLen(input)] = 0;
          *(double *)output = taosStr2Double(buf, NULL);
892
        } else if (inputType == TSDB_DATA_TYPE_NCHAR) {
H
Hongze Cheng 已提交
893
          int32_t len = taosUcs4ToMbs((TdUcs4 *)varDataVal(input), varDataLen(input), convBuf);
894
          if (len < 0) {
895 896
            code = TSDB_CODE_FAILED;
            goto _end;
897
          }
898 899
          convBuf[len] = 0;
          *(double *)output = taosStr2Double(convBuf, NULL);
900 901 902 903 904
        } else {
          GET_TYPED_DATA(*(double *)output, double, inputType, input);
        }
        break;
      }
G
Ganlin Zhao 已提交
905 906
      case TSDB_DATA_TYPE_BOOL: {
        if (inputType == TSDB_DATA_TYPE_BINARY) {
907 908 909
          memcpy(buf, varDataVal(input), varDataLen(input));
          buf[varDataLen(input)] = 0;
          *(bool *)output = taosStr2Int8(buf, NULL, 10);
G
Ganlin Zhao 已提交
910
        } else if (inputType == TSDB_DATA_TYPE_NCHAR) {
H
Hongze Cheng 已提交
911
          int32_t len = taosUcs4ToMbs((TdUcs4 *)varDataVal(input), varDataLen(input), convBuf);
G
Ganlin Zhao 已提交
912
          if (len < 0) {
913 914
            code = TSDB_CODE_FAILED;
            goto _end;
G
Ganlin Zhao 已提交
915
          }
916 917
          convBuf[len] = 0;
          *(bool *)output = taosStr2Int8(convBuf, NULL, 10);
G
Ganlin Zhao 已提交
918 919 920 921 922
        } else {
          GET_TYPED_DATA(*(bool *)output, bool, inputType, input);
        }
        break;
      }
G
Ganlin Zhao 已提交
923
      case TSDB_DATA_TYPE_TIMESTAMP: {
924
        int64_t timeVal;
G
Ganlin Zhao 已提交
925
        if (inputType == TSDB_DATA_TYPE_BINARY || inputType == TSDB_DATA_TYPE_NCHAR) {
926 927 928
          int64_t timePrec;
          GET_TYPED_DATA(timePrec, int64_t, GET_PARAM_TYPE(&pInput[1]), pInput[1].columnData->pData);
          int32_t ret = convertStringToTimestamp(inputType, input, timePrec, &timeVal);
929 930 931 932 933
          if (ret != TSDB_CODE_SUCCESS) {
            *(int64_t *)output = 0;
          } else {
            *(int64_t *)output = timeVal;
          }
G
Ganlin Zhao 已提交
934 935 936 937 938 939 940
        } else {
          GET_TYPED_DATA(*(int64_t *)output, int64_t, inputType, input);
        }
        break;
      }
      case TSDB_DATA_TYPE_BINARY: {
        if (inputType == TSDB_DATA_TYPE_BOOL) {
H
Haojun Liao 已提交
941
          // NOTE: sprintf will append '\0' at the end of string
H
Hongze Cheng 已提交
942 943
          int32_t len = sprintf(varDataVal(output), "%.*s", (int32_t)(outputLen - VARSTR_HEADER_SIZE),
                                *(int8_t *)input ? "true" : "false");
G
Ganlin Zhao 已提交
944 945
          varDataSetLen(output, len);
        } else if (inputType == TSDB_DATA_TYPE_BINARY) {
wafwerar's avatar
wafwerar 已提交
946
          int32_t len = TMIN(varDataLen(input), outputLen - VARSTR_HEADER_SIZE);
H
Haojun Liao 已提交
947
          memcpy(varDataVal(output), varDataVal(input), len);
G
Ganlin Zhao 已提交
948
          varDataSetLen(output, len);
949
        } else if (inputType == TSDB_DATA_TYPE_NCHAR) {
950
          int32_t len = taosUcs4ToMbs((TdUcs4 *)varDataVal(input), varDataLen(input), convBuf);
951
          if (len < 0) {
952 953
            code = TSDB_CODE_FAILED;
            goto _end;
954 955
          }
          len = TMIN(len, outputLen - VARSTR_HEADER_SIZE);
956
          memcpy(varDataVal(output), convBuf, len);
957
          varDataSetLen(output, len);
G
Ganlin Zhao 已提交
958
        } else {
959 960
          NUM_TO_STRING(inputType, input, sizeof(buf), buf);
          int32_t len = (int32_t)strlen(buf);
G
Ganlin Zhao 已提交
961
          len = (outputLen - VARSTR_HEADER_SIZE) > len ? len : (outputLen - VARSTR_HEADER_SIZE);
962
          memcpy(varDataVal(output), buf, len);
G
Ganlin Zhao 已提交
963 964 965 966 967 968
          varDataSetLen(output, len);
        }
        break;
      }
      case TSDB_DATA_TYPE_NCHAR: {
        int32_t outputCharLen = (outputLen - VARSTR_HEADER_SIZE) / TSDB_NCHAR_SIZE;
969
        int32_t len;
G
Ganlin Zhao 已提交
970 971
        if (inputType == TSDB_DATA_TYPE_BOOL) {
          char tmp[8] = {0};
H
Hongze Cheng 已提交
972
          len = sprintf(tmp, "%.*s", outputCharLen, *(int8_t *)input ? "true" : "false");
G
Ganlin Zhao 已提交
973 974
          bool ret = taosMbsToUcs4(tmp, len, (TdUcs4 *)varDataVal(output), outputLen - VARSTR_HEADER_SIZE, &len);
          if (!ret) {
975 976
            code = TSDB_CODE_FAILED;
            goto _end;
G
Ganlin Zhao 已提交
977
          }
978

G
Ganlin Zhao 已提交
979 980
          varDataSetLen(output, len);
        } else if (inputType == TSDB_DATA_TYPE_BINARY) {
981
          len = outputCharLen > varDataLen(input) ? varDataLen(input) : outputCharLen;
H
Hongze Cheng 已提交
982 983
          bool ret = taosMbsToUcs4(input + VARSTR_HEADER_SIZE, len, (TdUcs4 *)varDataVal(output),
                                   outputLen - VARSTR_HEADER_SIZE, &len);
G
Ganlin Zhao 已提交
984
          if (!ret) {
985 986
            code = TSDB_CODE_FAILED;
            goto _end;
G
Ganlin Zhao 已提交
987 988 989
          }
          varDataSetLen(output, len);
        } else if (inputType == TSDB_DATA_TYPE_NCHAR) {
990 991 992
          len = TMIN(outputLen - VARSTR_HEADER_SIZE, varDataLen(input));
          memcpy(output, input, len + VARSTR_HEADER_SIZE);
          varDataSetLen(output, len);
G
Ganlin Zhao 已提交
993
        } else {
994 995
          NUM_TO_STRING(inputType, input, sizeof(buf), buf);
          len = (int32_t)strlen(buf);
G
Ganlin Zhao 已提交
996
          len = outputCharLen > len ? len : outputCharLen;
997
          bool ret = taosMbsToUcs4(buf, len, (TdUcs4 *)varDataVal(output), outputLen - VARSTR_HEADER_SIZE, &len);
G
Ganlin Zhao 已提交
998
          if (!ret) {
999 1000
            code = TSDB_CODE_FAILED;
            goto _end;
G
Ganlin Zhao 已提交
1001 1002 1003
          }
          varDataSetLen(output, len);
        }
1004

H
Hongze Cheng 已提交
1005
        // for constant conversion, need to set proper length of pOutput description
G
Ganlin Zhao 已提交
1006
        if (len < outputLen) {
D
dapan1121 已提交
1007
          pOutput->columnData->info.bytes = len + VARSTR_HEADER_SIZE;
1008
        }
1009

G
Ganlin Zhao 已提交
1010 1011 1012
        break;
      }
      default: {
1013 1014
        code = TSDB_CODE_FAILED;
        goto _end;
G
Ganlin Zhao 已提交
1015 1016 1017 1018 1019 1020 1021
      }
    }

    colDataAppend(pOutput->columnData, i, output, false);
  }

  pOutput->numOfRows = pInput->numOfRows;
1022

H
Hongze Cheng 已提交
1023
_end:
1024
  taosMemoryFree(output);
H
Haojun Liao 已提交
1025
  taosMemoryFree(convBuf);
1026
  return code;
G
Ganlin Zhao 已提交
1027
}
G
Ganlin Zhao 已提交
1028

1029 1030 1031
int32_t toISO8601Function(SScalarParam *pInput, int32_t inputNum, SScalarParam *pOutput) {
  int32_t type = GET_PARAM_TYPE(pInput);

H
Hongze Cheng 已提交
1032
  bool    tzPresent = (inputNum == 2) ? true : false;
G
Ganlin Zhao 已提交
1033 1034
  char    tz[20] = {0};
  int32_t tzLen = 0;
1035 1036
  if (tzPresent) {
    tzLen = varDataLen(pInput[1].columnData->pData);
G
Ganlin Zhao 已提交
1037
    memcpy(tz, varDataVal(pInput[1].columnData->pData), tzLen);
1038
  }
1039

1040 1041 1042 1043 1044 1045
  for (int32_t i = 0; i < pInput[0].numOfRows; ++i) {
    if (colDataIsNull_s(pInput[0].columnData, i)) {
      colDataAppendNULL(pOutput->columnData, i);
      continue;
    }

1046
    char *input = colDataGetData(pInput[0].columnData, i);
H
Hongze Cheng 已提交
1047 1048
    char  fraction[20] = {0};
    bool  hasFraction = false;
1049 1050 1051
    NUM_TO_STRING(type, input, sizeof(fraction), fraction);
    int32_t tsDigits = (int32_t)strlen(fraction);

H
Hongze Cheng 已提交
1052
    char    buf[64] = {0};
1053 1054 1055 1056 1057 1058 1059 1060 1061 1062
    int64_t timeVal;
    GET_TYPED_DATA(timeVal, int64_t, type, input);
    if (tsDigits > TSDB_TIME_PRECISION_SEC_DIGITS) {
      if (tsDigits == TSDB_TIME_PRECISION_MILLI_DIGITS) {
        timeVal = timeVal / 1000;
      } else if (tsDigits == TSDB_TIME_PRECISION_MICRO_DIGITS) {
        timeVal = timeVal / (1000 * 1000);
      } else if (tsDigits == TSDB_TIME_PRECISION_NANO_DIGITS) {
        timeVal = timeVal / (1000 * 1000 * 1000);
      } else {
1063 1064
        colDataAppendNULL(pOutput->columnData, i);
        continue;
1065 1066 1067 1068 1069
      }
      hasFraction = true;
      memmove(fraction, fraction + TSDB_TIME_PRECISION_SEC_DIGITS, TSDB_TIME_PRECISION_SEC_DIGITS);
    }

1070 1071 1072
    struct tm tmInfo;
    taosLocalTime((const time_t *)&timeVal, &tmInfo);
    strftime(buf, sizeof(buf), "%Y-%m-%dT%H:%M:%S", &tmInfo);
1073 1074
    int32_t len = (int32_t)strlen(buf);

H
Hongze Cheng 已提交
1075
    // add timezone string
G
Ganlin Zhao 已提交
1076 1077 1078 1079
    if (tzLen > 0) {
      snprintf(buf + len, tzLen + 1, "%s", tz);
      len += tzLen;
    }
1080

1081 1082
    if (hasFraction) {
      int32_t fracLen = (int32_t)strlen(fraction) + 1;
1083 1084 1085 1086

      char *tzInfo;
      if (buf[len - 1] == 'z' || buf[len - 1] == 'Z') {
        tzInfo = &buf[len - 1];
1087 1088
        memmove(tzInfo + fracLen, tzInfo, strlen(tzInfo));
      } else {
1089 1090 1091 1092
        tzInfo = strchr(buf, '+');
        if (tzInfo) {
          memmove(tzInfo + fracLen, tzInfo, strlen(tzInfo));
        } else {
H
Hongze Cheng 已提交
1093
          // search '-' backwards
1094 1095 1096 1097 1098
          tzInfo = strrchr(buf, '-');
          if (tzInfo) {
            memmove(tzInfo + fracLen, tzInfo, strlen(tzInfo));
          }
        }
1099 1100
      }

1101
      char tmp[32] = {0};
1102 1103 1104 1105 1106 1107 1108 1109 1110 1111 1112 1113 1114 1115 1116 1117
      sprintf(tmp, ".%s", fraction);
      memcpy(tzInfo, tmp, fracLen);
      len += fracLen;
    }

    memmove(buf + VARSTR_HEADER_SIZE, buf, len);
    varDataSetLen(buf, len);

    colDataAppend(pOutput->columnData, i, buf, false);
  }

  pOutput->numOfRows = pInput->numOfRows;

  return TSDB_CODE_SUCCESS;
}

1118 1119
int32_t toUnixtimestampFunction(SScalarParam *pInput, int32_t inputNum, SScalarParam *pOutput) {
  int32_t type = GET_PARAM_TYPE(pInput);
1120 1121
  int64_t timePrec;
  GET_TYPED_DATA(timePrec, int64_t, GET_PARAM_TYPE(&pInput[1]), pInput[1].columnData->pData);
1122 1123 1124 1125 1126 1127

  for (int32_t i = 0; i < pInput[0].numOfRows; ++i) {
    if (colDataIsNull_s(pInput[0].columnData, i)) {
      colDataAppendNULL(pOutput->columnData, i);
      continue;
    }
1128
    char *input = colDataGetData(pInput[0].columnData, i);
1129 1130

    int64_t timeVal = 0;
1131 1132 1133
    int32_t ret = convertStringToTimestamp(type, input, timePrec, &timeVal);
    if (ret != TSDB_CODE_SUCCESS) {
      colDataAppendNULL(pOutput->columnData, i);
1134 1135
    } else {
      colDataAppend(pOutput->columnData, i, (char *)&timeVal, false);
1136
    }
1137 1138 1139 1140 1141 1142 1143
  }

  pOutput->numOfRows = pInput->numOfRows;

  return TSDB_CODE_SUCCESS;
}

wmmhello's avatar
wmmhello 已提交
1144 1145 1146
int32_t toJsonFunction(SScalarParam *pInput, int32_t inputNum, SScalarParam *pOutput) {
  int32_t type = GET_PARAM_TYPE(pInput);

wmmhello's avatar
wmmhello 已提交
1147
  char tmp[TSDB_MAX_JSON_TAG_LEN] = {0};
wmmhello's avatar
wmmhello 已提交
1148
  for (int32_t i = 0; i < pInput[0].numOfRows; ++i) {
H
Hongze Cheng 已提交
1149 1150
    SArray *pTagVals = taosArrayInit(8, sizeof(STagVal));
    STag   *pTag = NULL;
wmmhello's avatar
wmmhello 已提交
1151

wmmhello's avatar
wmmhello 已提交
1152 1153
    if (colDataIsNull_s(pInput[0].columnData, i)) {
      tTagNew(pTagVals, 1, true, &pTag);
H
Hongze Cheng 已提交
1154
    } else {
wmmhello's avatar
wmmhello 已提交
1155
      char *input = pInput[0].columnData->pData + pInput[0].columnData->varmeta.offset[i];
H
Hongze Cheng 已提交
1156
      if (varDataLen(input) > (TSDB_MAX_JSON_TAG_LEN - VARSTR_HEADER_SIZE) / TSDB_NCHAR_SIZE) {
wmmhello's avatar
wmmhello 已提交
1157 1158
        taosArrayDestroy(pTagVals);
        return TSDB_CODE_FAILED;
wmmhello's avatar
wmmhello 已提交
1159 1160
      }
      memcpy(tmp, varDataVal(input), varDataLen(input));
wmmhello's avatar
wmmhello 已提交
1161
      tmp[varDataLen(input)] = 0;
H
Hongze Cheng 已提交
1162
      if (parseJsontoTagData(tmp, pTagVals, &pTag, NULL)) {
wmmhello's avatar
wmmhello 已提交
1163 1164
        tTagNew(pTagVals, 1, true, &pTag);
      }
wmmhello's avatar
wmmhello 已提交
1165 1166
    }

H
Hongze Cheng 已提交
1167
    colDataAppend(pOutput->columnData, i, (const char *)pTag, false);
wmmhello's avatar
wmmhello 已提交
1168 1169
    tTagFree(pTag);
    taosArrayDestroy(pTagVals);
wmmhello's avatar
wmmhello 已提交
1170 1171 1172 1173 1174 1175
  }

  pOutput->numOfRows = pInput->numOfRows;
  return TSDB_CODE_SUCCESS;
}

1176
/** Time functions **/
1177 1178 1179
int32_t timeTruncateFunction(SScalarParam *pInput, int32_t inputNum, SScalarParam *pOutput) {
  int32_t type = GET_PARAM_TYPE(&pInput[0]);

1180
  int64_t timeUnit, timePrec, timeVal = 0;
1181
  GET_TYPED_DATA(timeUnit, int64_t, GET_PARAM_TYPE(&pInput[1]), pInput[1].columnData->pData);
1182
  GET_TYPED_DATA(timePrec, int64_t, GET_PARAM_TYPE(&pInput[2]), pInput[2].columnData->pData);
1183

1184
  int64_t factor = TSDB_TICK_PER_SECOND(timePrec);
1185
  int64_t unit = timeUnit * 1000 / factor;
1186

1187 1188 1189 1190 1191 1192
  for (int32_t i = 0; i < pInput[0].numOfRows; ++i) {
    if (colDataIsNull_s(pInput[0].columnData, i)) {
      colDataAppendNULL(pOutput->columnData, i);
      continue;
    }

1193 1194
    char *input = colDataGetData(pInput[0].columnData, i);

1195
    if (IS_VAR_DATA_TYPE(type)) { /* datetime format strings */
1196 1197 1198 1199 1200
      int32_t ret = convertStringToTimestamp(type, input, TSDB_TIME_PRECISION_NANO, &timeVal);
      if (ret != TSDB_CODE_SUCCESS) {
        colDataAppendNULL(pOutput->columnData, i);
        continue;
      }
H
Hongze Cheng 已提交
1201
      // If converted value is less than 10digits in second, use value in second instead
1202 1203 1204 1205 1206 1207 1208 1209 1210 1211 1212 1213 1214 1215 1216 1217 1218 1219
      int64_t timeValSec = timeVal / 1000000000;
      if (timeValSec < 1000000000) {
        timeVal = timeValSec;
      }
    } else if (type == TSDB_DATA_TYPE_BIGINT) { /* unix timestamp */
      GET_TYPED_DATA(timeVal, int64_t, type, input);
    } else if (type == TSDB_DATA_TYPE_TIMESTAMP) { /* timestamp column*/
      GET_TYPED_DATA(timeVal, int64_t, type, input);
      int64_t timeValSec = timeVal / factor;
      if (timeValSec < 1000000000) {
        timeVal = timeValSec;
      }
    }

    char buf[20] = {0};
    NUM_TO_STRING(TSDB_DATA_TYPE_BIGINT, &timeVal, sizeof(buf), buf);
    int32_t tsDigits = (int32_t)strlen(buf);

1220 1221
    switch (unit) {
      case 0: { /* 1u or 1b */
1222
        if (tsDigits == TSDB_TIME_PRECISION_NANO_DIGITS) {
1223 1224 1225 1226 1227
          if (timePrec == TSDB_TIME_PRECISION_NANO && timeUnit == 1) {
            timeVal = timeVal * 1;
          } else {
            timeVal = timeVal / 1000 * 1000;
          }
1228 1229 1230 1231 1232 1233 1234 1235 1236 1237 1238 1239 1240 1241
        } else if (tsDigits <= TSDB_TIME_PRECISION_SEC_DIGITS) {
          timeVal = timeVal * factor;
        } else {
          timeVal = timeVal * 1;
        }
        break;
      }
      case 1: { /* 1a */
        if (tsDigits == TSDB_TIME_PRECISION_MILLI_DIGITS) {
          timeVal = timeVal * 1;
        } else if (tsDigits == TSDB_TIME_PRECISION_MICRO_DIGITS) {
          timeVal = timeVal / 1000 * 1000;
        } else if (tsDigits == TSDB_TIME_PRECISION_NANO_DIGITS) {
          timeVal = timeVal / 1000000 * 1000000;
H
Hongze Cheng 已提交
1242
        } else if (tsDigits <= TSDB_TIME_PRECISION_SEC_DIGITS) {
1243 1244
          timeVal = timeVal * factor;
        } else {
1245 1246
          colDataAppendNULL(pOutput->columnData, i);
          continue;
1247 1248 1249 1250 1251 1252 1253 1254 1255 1256 1257 1258 1259
        }
        break;
      }
      case 1000: { /* 1s */
        if (tsDigits == TSDB_TIME_PRECISION_MILLI_DIGITS) {
          timeVal = timeVal / 1000 * 1000;
        } else if (tsDigits == TSDB_TIME_PRECISION_MICRO_DIGITS) {
          timeVal = timeVal / 1000000 * 1000000;
        } else if (tsDigits == TSDB_TIME_PRECISION_NANO_DIGITS) {
          timeVal = timeVal / 1000000000 * 1000000000;
        } else if (tsDigits <= TSDB_TIME_PRECISION_SEC_DIGITS) {
          timeVal = timeVal * factor;
        } else {
1260 1261
          colDataAppendNULL(pOutput->columnData, i);
          continue;
1262 1263 1264 1265 1266 1267 1268 1269 1270 1271 1272 1273 1274
        }
        break;
      }
      case 60000: { /* 1m */
        if (tsDigits == TSDB_TIME_PRECISION_MILLI_DIGITS) {
          timeVal = timeVal / 1000 / 60 * 60 * 1000;
        } else if (tsDigits == TSDB_TIME_PRECISION_MICRO_DIGITS) {
          timeVal = timeVal / 1000000 / 60 * 60 * 1000000;
        } else if (tsDigits == TSDB_TIME_PRECISION_NANO_DIGITS) {
          timeVal = timeVal / 1000000000 / 60 * 60 * 1000000000;
        } else if (tsDigits <= TSDB_TIME_PRECISION_SEC_DIGITS) {
          timeVal = timeVal * factor / factor / 60 * 60 * factor;
        } else {
1275 1276
          colDataAppendNULL(pOutput->columnData, i);
          continue;
1277 1278 1279 1280 1281 1282 1283 1284 1285 1286 1287 1288 1289
        }
        break;
      }
      case 3600000: { /* 1h */
        if (tsDigits == TSDB_TIME_PRECISION_MILLI_DIGITS) {
          timeVal = timeVal / 1000 / 3600 * 3600 * 1000;
        } else if (tsDigits == TSDB_TIME_PRECISION_MICRO_DIGITS) {
          timeVal = timeVal / 1000000 / 3600 * 3600 * 1000000;
        } else if (tsDigits == TSDB_TIME_PRECISION_NANO_DIGITS) {
          timeVal = timeVal / 1000000000 / 3600 * 3600 * 1000000000;
        } else if (tsDigits <= TSDB_TIME_PRECISION_SEC_DIGITS) {
          timeVal = timeVal * factor / factor / 3600 * 3600 * factor;
        } else {
1290 1291
          colDataAppendNULL(pOutput->columnData, i);
          continue;
1292 1293 1294 1295 1296 1297 1298 1299 1300 1301 1302
        }
        break;
      }
      case 86400000: { /* 1d */
        if (tsDigits == TSDB_TIME_PRECISION_MILLI_DIGITS) {
          timeVal = timeVal / 1000 / 86400 * 86400 * 1000;
        } else if (tsDigits == TSDB_TIME_PRECISION_MICRO_DIGITS) {
          timeVal = timeVal / 1000000 / 86400 * 86400 * 1000000;
        } else if (tsDigits == TSDB_TIME_PRECISION_NANO_DIGITS) {
          timeVal = timeVal / 1000000000 / 86400 * 86400 * 1000000000;
        } else if (tsDigits <= TSDB_TIME_PRECISION_SEC_DIGITS) {
H
Hongze Cheng 已提交
1303
          timeVal = timeVal * factor / factor / 86400 * 86400 * factor;
1304
        } else {
1305 1306
          colDataAppendNULL(pOutput->columnData, i);
          continue;
1307 1308 1309 1310 1311 1312 1313 1314 1315 1316 1317
        }
        break;
      }
      case 604800000: { /* 1w */
        if (tsDigits == TSDB_TIME_PRECISION_MILLI_DIGITS) {
          timeVal = timeVal / 1000 / 604800 * 604800 * 1000;
        } else if (tsDigits == TSDB_TIME_PRECISION_MICRO_DIGITS) {
          timeVal = timeVal / 1000000 / 604800 * 604800 * 1000000;
        } else if (tsDigits == TSDB_TIME_PRECISION_NANO_DIGITS) {
          timeVal = timeVal / 1000000000 / 604800 * 604800 * 1000000000;
        } else if (tsDigits <= TSDB_TIME_PRECISION_SEC_DIGITS) {
H
Hongze Cheng 已提交
1318
          timeVal = timeVal * factor / factor / 604800 * 604800 * factor;
1319
        } else {
1320 1321
          colDataAppendNULL(pOutput->columnData, i);
          continue;
1322 1323 1324 1325 1326 1327 1328 1329 1330
        }
        break;
      }
      default: {
        timeVal = timeVal * 1;
        break;
      }
    }

H
Hongze Cheng 已提交
1331
    // truncate the timestamp to db precision
1332 1333 1334 1335 1336 1337 1338 1339 1340 1341 1342 1343 1344 1345 1346 1347 1348 1349 1350 1351 1352 1353 1354 1355 1356 1357 1358 1359 1360 1361 1362 1363 1364 1365 1366
    switch (timePrec) {
      case TSDB_TIME_PRECISION_MILLI: {
        if (tsDigits == TSDB_TIME_PRECISION_MICRO_DIGITS) {
          timeVal = timeVal / 1000;
        } else if (tsDigits == TSDB_TIME_PRECISION_NANO_DIGITS) {
          timeVal = timeVal / 1000000;
        }
        break;
      }
      case TSDB_TIME_PRECISION_MICRO: {
        if (tsDigits == TSDB_TIME_PRECISION_NANO_DIGITS) {
          timeVal = timeVal / 1000;
        } else if (tsDigits == TSDB_TIME_PRECISION_MILLI_DIGITS) {
          timeVal = timeVal * 1000;
        }
        break;
      }
      case TSDB_TIME_PRECISION_NANO: {
        if (tsDigits == TSDB_TIME_PRECISION_MICRO_DIGITS) {
          timeVal = timeVal * 1000;
        } else if (tsDigits == TSDB_TIME_PRECISION_MILLI_DIGITS) {
          timeVal = timeVal * 1000000;
        }
        break;
      }
    }

    colDataAppend(pOutput->columnData, i, (char *)&timeVal, false);
  }

  pOutput->numOfRows = pInput->numOfRows;

  return TSDB_CODE_SUCCESS;
}

G
Ganlin Zhao 已提交
1367
int32_t timeDiffFunction(SScalarParam *pInput, int32_t inputNum, SScalarParam *pOutput) {
1368 1369
  int64_t timeUnit = -1, timePrec, timeVal[2] = {0};
  if (inputNum == 4) {
G
Ganlin Zhao 已提交
1370
    GET_TYPED_DATA(timeUnit, int64_t, GET_PARAM_TYPE(&pInput[2]), pInput[2].columnData->pData);
1371 1372 1373
    GET_TYPED_DATA(timePrec, int64_t, GET_PARAM_TYPE(&pInput[3]), pInput[3].columnData->pData);
  } else {
    GET_TYPED_DATA(timePrec, int64_t, GET_PARAM_TYPE(&pInput[2]), pInput[2].columnData->pData);
G
Ganlin Zhao 已提交
1374 1375
  }

1376
  int64_t factor = TSDB_TICK_PER_SECOND(timePrec);
G
Ganlin Zhao 已提交
1377 1378 1379 1380
  int32_t numOfRows = 0;
  for (int32_t i = 0; i < inputNum; ++i) {
    if (pInput[i].numOfRows > numOfRows) {
      numOfRows = pInput[i].numOfRows;
G
Ganlin Zhao 已提交
1381 1382 1383
    }
  }

G
Ganlin Zhao 已提交
1384 1385 1386
  char *input[2];
  for (int32_t i = 0; i < numOfRows; ++i) {
    bool hasNull = false;
G
Ganlin Zhao 已提交
1387
    for (int32_t k = 0; k < 2; ++k) {
G
Ganlin Zhao 已提交
1388 1389 1390
      if (colDataIsNull_s(pInput[k].columnData, i)) {
        hasNull = true;
        break;
G
Ganlin Zhao 已提交
1391 1392
      }

1393 1394 1395
      int32_t rowIdx = (pInput[k].numOfRows == 1) ? 0 : i;
      input[k] = colDataGetData(pInput[k].columnData, rowIdx);

G
Ganlin Zhao 已提交
1396 1397
      int32_t type = GET_PARAM_TYPE(&pInput[k]);
      if (IS_VAR_DATA_TYPE(type)) { /* datetime format strings */
1398 1399
        int32_t ret = convertStringToTimestamp(type, input[k], TSDB_TIME_PRECISION_NANO, &timeVal[k]);
        if (ret != TSDB_CODE_SUCCESS) {
1400 1401
          hasNull = true;
          break;
1402
        }
G
Ganlin Zhao 已提交
1403 1404 1405 1406 1407 1408 1409 1410 1411 1412 1413 1414 1415 1416 1417 1418 1419 1420 1421
      } else if (type == TSDB_DATA_TYPE_BIGINT || type == TSDB_DATA_TYPE_TIMESTAMP) { /* unix timestamp or ts column*/
        GET_TYPED_DATA(timeVal[k], int64_t, type, input[k]);
        if (type == TSDB_DATA_TYPE_TIMESTAMP) {
          int64_t timeValSec = timeVal[k] / factor;
          if (timeValSec < 1000000000) {
            timeVal[k] = timeValSec;
          }
        }

        char buf[20] = {0};
        NUM_TO_STRING(TSDB_DATA_TYPE_BIGINT, &timeVal[k], sizeof(buf), buf);
        int32_t tsDigits = (int32_t)strlen(buf);
        if (tsDigits <= TSDB_TIME_PRECISION_SEC_DIGITS) {
          timeVal[k] = timeVal[k] * 1000000000;
        } else if (tsDigits == TSDB_TIME_PRECISION_MILLI_DIGITS) {
          timeVal[k] = timeVal[k] * 1000000;
        } else if (tsDigits == TSDB_TIME_PRECISION_MICRO_DIGITS) {
          timeVal[k] = timeVal[k] * 1000;
        } else if (tsDigits == TSDB_TIME_PRECISION_NANO_DIGITS) {
G
Ganlin Zhao 已提交
1422
          timeVal[k] = timeVal[k] * 1;
G
Ganlin Zhao 已提交
1423
        } else {
1424 1425
          hasNull = true;
          break;
G
Ganlin Zhao 已提交
1426
        }
G
Ganlin Zhao 已提交
1427 1428 1429
      }
    }

G
Ganlin Zhao 已提交
1430 1431 1432 1433 1434
    if (hasNull) {
      colDataAppendNULL(pOutput->columnData, i);
      continue;
    }

H
Hongze Cheng 已提交
1435
    int64_t result = (timeVal[0] >= timeVal[1]) ? (timeVal[0] - timeVal[1]) : (timeVal[1] - timeVal[0]);
G
Ganlin Zhao 已提交
1436

H
Hongze Cheng 已提交
1437 1438
    if (timeUnit < 0) {  // if no time unit given use db precision
      switch (timePrec) {
G
Ganlin Zhao 已提交
1439 1440 1441 1442 1443 1444 1445 1446 1447 1448 1449 1450 1451 1452
        case TSDB_TIME_PRECISION_MILLI: {
          result = result / 1000000;
          break;
        }
        case TSDB_TIME_PRECISION_MICRO: {
          result = result / 1000;
          break;
        }
        case TSDB_TIME_PRECISION_NANO: {
          result = result / 1;
          break;
        }
      }
    } else {
1453
      int64_t unit = timeUnit * 1000 / factor;
H
Hongze Cheng 已提交
1454
      switch (unit) {
1455 1456 1457 1458 1459 1460
        case 0: { /* 1u or 1b */
          if (timePrec == TSDB_TIME_PRECISION_NANO && timeUnit == 1) {
            result = result / 1;
          } else {
            result = result / 1000;
          }
G
Ganlin Zhao 已提交
1461 1462 1463 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
          break;
        }
        case 1: { /* 1a */
          result = result / 1000000;
          break;
        }
        case 1000: { /* 1s */
          result = result / 1000000000;
          break;
        }
        case 60000: { /* 1m */
          result = result / 1000000000 / 60;
          break;
        }
        case 3600000: { /* 1h */
          result = result / 1000000000 / 3600;
          break;
        }
        case 86400000: { /* 1d */
          result = result / 1000000000 / 86400;
          break;
        }
        case 604800000: { /* 1w */
          result = result / 1000000000 / 604800;
          break;
        }
        default: {
          break;
        }
      }
    }

    colDataAppend(pOutput->columnData, i, (char *)&result, false);
  }

G
Ganlin Zhao 已提交
1496
  pOutput->numOfRows = numOfRows;
G
Ganlin Zhao 已提交
1497 1498 1499 1500

  return TSDB_CODE_SUCCESS;
}

1501
int32_t nowFunction(SScalarParam *pInput, int32_t inputNum, SScalarParam *pOutput) {
1502 1503 1504 1505
  int64_t timePrec;
  GET_TYPED_DATA(timePrec, int64_t, GET_PARAM_TYPE(&pInput[0]), pInput[0].columnData->pData);

  int64_t ts = taosGetTimestamp(timePrec);
X
Xiaoyu Wang 已提交
1506 1507
  for (int32_t i = 0; i < pInput->numOfRows; ++i) {
    colDataAppendInt64(pOutput->columnData, i, &ts);
1508
  }
X
Xiaoyu Wang 已提交
1509
  pOutput->numOfRows = pInput->numOfRows;
1510 1511 1512 1513
  return TSDB_CODE_SUCCESS;
}

int32_t todayFunction(SScalarParam *pInput, int32_t inputNum, SScalarParam *pOutput) {
1514
  int64_t timePrec;
1515
  GET_TYPED_DATA(timePrec, int64_t, GET_PARAM_TYPE(&pInput[0]), pInput[0].columnData->pData);
1516 1517

  int64_t ts = taosGetTimestampToday(timePrec);
X
Xiaoyu Wang 已提交
1518 1519
  for (int32_t i = 0; i < pInput->numOfRows; ++i) {
    colDataAppendInt64(pOutput->columnData, i, &ts);
1520
  }
X
Xiaoyu Wang 已提交
1521
  pOutput->numOfRows = pInput->numOfRows;
1522 1523 1524
  return TSDB_CODE_SUCCESS;
}

1525
int32_t timezoneFunction(SScalarParam *pInput, int32_t inputNum, SScalarParam *pOutput) {
1526 1527 1528
  char output[TD_TIMEZONE_LEN + VARSTR_HEADER_SIZE] = {0};
  memcpy(varDataVal(output), tsTimezoneStr, TD_TIMEZONE_LEN);
  varDataSetLen(output, strlen(tsTimezoneStr));
X
Xiaoyu Wang 已提交
1529
  for (int32_t i = 0; i < pInput->numOfRows; ++i) {
1530
    colDataAppend(pOutput->columnData, i, output, false);
1531
  }
X
Xiaoyu Wang 已提交
1532
  pOutput->numOfRows = pInput->numOfRows;
1533 1534 1535
  return TSDB_CODE_SUCCESS;
}

1536 1537 1538
int32_t atanFunction(SScalarParam *pInput, int32_t inputNum, SScalarParam *pOutput) {
  return doScalarFunctionUnique(pInput, inputNum, pOutput, atan);
}
1539

1540 1541 1542
int32_t sinFunction(SScalarParam *pInput, int32_t inputNum, SScalarParam *pOutput) {
  return doScalarFunctionUnique(pInput, inputNum, pOutput, sin);
}
1543

1544 1545 1546
int32_t cosFunction(SScalarParam *pInput, int32_t inputNum, SScalarParam *pOutput) {
  return doScalarFunctionUnique(pInput, inputNum, pOutput, cos);
}
1547

1548 1549 1550
int32_t tanFunction(SScalarParam *pInput, int32_t inputNum, SScalarParam *pOutput) {
  return doScalarFunctionUnique(pInput, inputNum, pOutput, tan);
}
1551

1552 1553 1554
int32_t asinFunction(SScalarParam *pInput, int32_t inputNum, SScalarParam *pOutput) {
  return doScalarFunctionUnique(pInput, inputNum, pOutput, asin);
}
1555

1556 1557 1558
int32_t acosFunction(SScalarParam *pInput, int32_t inputNum, SScalarParam *pOutput) {
  return doScalarFunctionUnique(pInput, inputNum, pOutput, acos);
}
H
Haojun Liao 已提交
1559

G
Ganlin Zhao 已提交
1560 1561 1562 1563 1564
int32_t powFunction(SScalarParam *pInput, int32_t inputNum, SScalarParam *pOutput) {
  return doScalarFunctionUnique2(pInput, inputNum, pOutput, pow);
}

int32_t logFunction(SScalarParam *pInput, int32_t inputNum, SScalarParam *pOutput) {
1565 1566 1567 1568 1569
  if (inputNum == 1) {
    return doScalarFunctionUnique(pInput, inputNum, pOutput, tlog);
  } else {
    return doScalarFunctionUnique2(pInput, inputNum, pOutput, tlog2);
  }
G
Ganlin Zhao 已提交
1570 1571
}

1572 1573 1574
int32_t sqrtFunction(SScalarParam *pInput, int32_t inputNum, SScalarParam *pOutput) {
  return doScalarFunctionUnique(pInput, inputNum, pOutput, sqrt);
}
1575

1576 1577 1578 1579 1580 1581 1582 1583 1584 1585
int32_t ceilFunction(SScalarParam *pInput, int32_t inputNum, SScalarParam *pOutput) {
  return doScalarFunction(pInput, inputNum, pOutput, ceilf, ceil);
}

int32_t floorFunction(SScalarParam *pInput, int32_t inputNum, SScalarParam *pOutput) {
  return doScalarFunction(pInput, inputNum, pOutput, floorf, floor);
}

int32_t roundFunction(SScalarParam *pInput, int32_t inputNum, SScalarParam *pOutput) {
  return doScalarFunction(pInput, inputNum, pOutput, roundf, round);
1586 1587
}

G
Ganlin Zhao 已提交
1588
int32_t lowerFunction(SScalarParam *pInput, int32_t inputNum, SScalarParam *pOutput) {
wafwerar's avatar
wafwerar 已提交
1589 1590 1591
#ifdef WINDOWS
  return doCaseConvFunction(pInput, inputNum, pOutput, towlower);
#else
G
Ganlin Zhao 已提交
1592
  return doCaseConvFunction(pInput, inputNum, pOutput, tolower);
wafwerar's avatar
wafwerar 已提交
1593
#endif
1594 1595
}

G
Ganlin Zhao 已提交
1596
int32_t upperFunction(SScalarParam *pInput, int32_t inputNum, SScalarParam *pOutput) {
wafwerar's avatar
wafwerar 已提交
1597 1598 1599
#ifdef WINDOWS
  return doCaseConvFunction(pInput, inputNum, pOutput, towupper);
#else
G
Ganlin Zhao 已提交
1600
  return doCaseConvFunction(pInput, inputNum, pOutput, toupper);
wafwerar's avatar
wafwerar 已提交
1601
#endif
1602 1603
}

G
Ganlin Zhao 已提交
1604 1605
int32_t ltrimFunction(SScalarParam *pInput, int32_t inputNum, SScalarParam *pOutput) {
  return doTrimFunction(pInput, inputNum, pOutput, tltrim);
1606 1607
}

G
Ganlin Zhao 已提交
1608 1609
int32_t rtrimFunction(SScalarParam *pInput, int32_t inputNum, SScalarParam *pOutput) {
  return doTrimFunction(pInput, inputNum, pOutput, trtrim);
1610 1611
}

G
Ganlin Zhao 已提交
1612 1613 1614
int32_t lengthFunction(SScalarParam *pInput, int32_t inputNum, SScalarParam *pOutput) {
  return doLengthFunction(pInput, inputNum, pOutput, tlength);
}
1615

G
Ganlin Zhao 已提交
1616 1617
int32_t charLengthFunction(SScalarParam *pInput, int32_t inputNum, SScalarParam *pOutput) {
  return doLengthFunction(pInput, inputNum, pOutput, tcharlength);
1618 1619
}

H
Haojun Liao 已提交
1620
#if 0
1621 1622 1623 1624 1625 1626 1627 1628 1629 1630 1631 1632 1633 1634 1635 1636 1637 1638 1639 1640 1641 1642 1643 1644 1645 1646 1647 1648 1649 1650 1651 1652 1653 1654 1655 1656 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
static void reverseCopy(char* dest, const char* src, int16_t type, int32_t numOfRows) {
  switch(type) {
    case TSDB_DATA_TYPE_TINYINT:
    case TSDB_DATA_TYPE_UTINYINT:{
      int8_t* p = (int8_t*) dest;
      int8_t* pSrc = (int8_t*) src;

      for(int32_t i = 0; i < numOfRows; ++i) {
        p[i] = pSrc[numOfRows - i - 1];
      }
      return;
    }

    case TSDB_DATA_TYPE_SMALLINT:
    case TSDB_DATA_TYPE_USMALLINT:{
      int16_t* p = (int16_t*) dest;
      int16_t* pSrc = (int16_t*) src;

      for(int32_t i = 0; i < numOfRows; ++i) {
        p[i] = pSrc[numOfRows - i - 1];
      }
      return;
    }
    case TSDB_DATA_TYPE_INT:
    case TSDB_DATA_TYPE_UINT: {
      int32_t* p = (int32_t*) dest;
      int32_t* pSrc = (int32_t*) src;

      for(int32_t i = 0; i < numOfRows; ++i) {
        p[i] = pSrc[numOfRows - i - 1];
      }
      return;
    }
    case TSDB_DATA_TYPE_BIGINT:
    case TSDB_DATA_TYPE_UBIGINT: {
      int64_t* p = (int64_t*) dest;
      int64_t* pSrc = (int64_t*) src;

      for(int32_t i = 0; i < numOfRows; ++i) {
        p[i] = pSrc[numOfRows - i - 1];
      }
      return;
    }
    case TSDB_DATA_TYPE_FLOAT: {
      float* p = (float*) dest;
      float* pSrc = (float*) src;

      for(int32_t i = 0; i < numOfRows; ++i) {
        p[i] = pSrc[numOfRows - i - 1];
      }
      return;
    }
    case TSDB_DATA_TYPE_DOUBLE: {
      double* p = (double*) dest;
      double* pSrc = (double*) src;

      for(int32_t i = 0; i < numOfRows; ++i) {
        p[i] = pSrc[numOfRows - i - 1];
      }
      return;
    }
    default: assert(0);
  }
}
H
Haojun Liao 已提交
1685
#endif
1686

H
Hongze Cheng 已提交
1687
bool getTimePseudoFuncEnv(SFunctionNode *UNUSED_PARAM(pFunc), SFuncExecEnv *pEnv) {
1688 1689 1690 1691 1692 1693 1694
  pEnv->calcMemSize = sizeof(int64_t);
  return true;
}

int32_t qStartTsFunction(SScalarParam *pInput, int32_t inputNum, SScalarParam *pOutput) {
  ASSERT(inputNum == 1);
  colDataAppendInt64(pOutput->columnData, pOutput->numOfRows, (int64_t *)colDataGetData(pInput->columnData, 0));
1695
  return TSDB_CODE_SUCCESS;
1696 1697 1698 1699 1700
}

int32_t qEndTsFunction(SScalarParam *pInput, int32_t inputNum, SScalarParam *pOutput) {
  ASSERT(inputNum == 1);
  colDataAppendInt64(pOutput->columnData, pOutput->numOfRows, (int64_t *)colDataGetData(pInput->columnData, 1));
1701
  return TSDB_CODE_SUCCESS;
1702 1703 1704 1705 1706
}

int32_t winDurFunction(SScalarParam *pInput, int32_t inputNum, SScalarParam *pOutput) {
  ASSERT(inputNum == 1);
  colDataAppendInt64(pOutput->columnData, pOutput->numOfRows, (int64_t *)colDataGetData(pInput->columnData, 2));
1707
  return TSDB_CODE_SUCCESS;
1708 1709 1710 1711
}

int32_t winStartTsFunction(SScalarParam *pInput, int32_t inputNum, SScalarParam *pOutput) {
  ASSERT(inputNum == 1);
H
Hongze Cheng 已提交
1712
  colDataAppendInt64(pOutput->columnData, pOutput->numOfRows, (int64_t *)colDataGetData(pInput->columnData, 3));
1713 1714 1715 1716 1717
  return TSDB_CODE_SUCCESS;
}

int32_t winEndTsFunction(SScalarParam *pInput, int32_t inputNum, SScalarParam *pOutput) {
  ASSERT(inputNum == 1);
H
Hongze Cheng 已提交
1718
  colDataAppendInt64(pOutput->columnData, pOutput->numOfRows, (int64_t *)colDataGetData(pInput->columnData, 4));
1719
  return TSDB_CODE_SUCCESS;
1720
}
1721 1722 1723

int32_t qTbnameFunction(SScalarParam *pInput, int32_t inputNum, SScalarParam *pOutput) {
  ASSERT(inputNum == 1);
H
Haojun Liao 已提交
1724
  char* p = colDataGetVarData(pInput->columnData, 0);
1725

H
Haojun Liao 已提交
1726
  colDataAppendNItems(pOutput->columnData, pOutput->numOfRows, p, pInput->numOfRows);
1727
  pOutput->numOfRows += pInput->numOfRows;
1728 1729
  return TSDB_CODE_SUCCESS;
}
1730 1731 1732

/** Aggregation functions **/
int32_t countScalarFunction(SScalarParam *pInput, int32_t inputNum, SScalarParam *pOutput) {
H
Hongze Cheng 已提交
1733
  SColumnInfoData *pInputData = pInput->columnData;
1734 1735 1736 1737 1738 1739
  SColumnInfoData *pOutputData = pOutput->columnData;

  int64_t *out = (int64_t *)pOutputData->pData;
  *out = 0;
  for (int32_t i = 0; i < pInput->numOfRows; ++i) {
    if (colDataIsNull_s(pInputData, i)) {
G
Ganlin Zhao 已提交
1740
      continue;
1741 1742 1743 1744
    }
    (*out)++;
  }

G
Ganlin Zhao 已提交
1745
  pOutput->numOfRows = 1;
1746 1747
  return TSDB_CODE_SUCCESS;
}
1748 1749

int32_t sumScalarFunction(SScalarParam *pInput, int32_t inputNum, SScalarParam *pOutput) {
H
Hongze Cheng 已提交
1750
  SColumnInfoData *pInputData = pInput->columnData;
1751 1752 1753
  SColumnInfoData *pOutputData = pOutput->columnData;

  int32_t type = GET_PARAM_TYPE(pInput);
H
Hongze Cheng 已提交
1754
  bool    hasNull = false;
1755 1756
  for (int32_t i = 0; i < pInput->numOfRows; ++i) {
    if (colDataIsNull_s(pInputData, i)) {
G
Ganlin Zhao 已提交
1757
      hasNull = true;
1758 1759 1760
      break;
    }

1761
    if (IS_SIGNED_NUMERIC_TYPE(type) || type == TSDB_DATA_TYPE_BOOL) {
1762
      int64_t *out = (int64_t *)pOutputData->pData;
1763 1764 1765 1766 1767 1768 1769 1770 1771 1772 1773 1774 1775
      if (type == TSDB_DATA_TYPE_TINYINT || type == TSDB_DATA_TYPE_BOOL) {
        int8_t *in = (int8_t *)pInputData->pData;
        *out += in[i];
      } else if (type == TSDB_DATA_TYPE_SMALLINT) {
        int16_t *in = (int16_t *)pInputData->pData;
        *out += in[i];
      } else if (type == TSDB_DATA_TYPE_INT) {
        int32_t *in = (int32_t *)pInputData->pData;
        *out += in[i];
      } else if (type == TSDB_DATA_TYPE_BIGINT) {
        int64_t *in = (int64_t *)pInputData->pData;
        *out += in[i];
      }
1776 1777
    } else if (IS_UNSIGNED_NUMERIC_TYPE(type)) {
      uint64_t *out = (uint64_t *)pOutputData->pData;
1778 1779 1780 1781 1782 1783 1784 1785 1786 1787 1788 1789 1790
      if (type == TSDB_DATA_TYPE_UTINYINT) {
        uint8_t *in = (uint8_t *)pInputData->pData;
        *out += in[i];
      } else if (type == TSDB_DATA_TYPE_USMALLINT) {
        uint16_t *in = (uint16_t *)pInputData->pData;
        *out += in[i];
      } else if (type == TSDB_DATA_TYPE_UINT) {
        uint32_t *in = (uint32_t *)pInputData->pData;
        *out += in[i];
      } else if (type == TSDB_DATA_TYPE_UBIGINT) {
        uint64_t *in = (uint64_t *)pInputData->pData;
        *out += in[i];
      }
1791 1792
    } else if (IS_FLOAT_TYPE(type)) {
      double *out = (double *)pOutputData->pData;
1793 1794 1795 1796 1797 1798 1799
      if (type == TSDB_DATA_TYPE_FLOAT) {
        float *in = (float *)pInputData->pData;
        *out += in[i];
      } else if (type == TSDB_DATA_TYPE_DOUBLE) {
        double *in = (double *)pInputData->pData;
        *out += in[i];
      }
1800 1801 1802
    }
  }

G
Ganlin Zhao 已提交
1803 1804 1805 1806 1807
  if (hasNull) {
    colDataAppendNULL(pOutputData, 0);
  }

  pOutput->numOfRows = 1;
1808 1809
  return TSDB_CODE_SUCCESS;
}
1810

1811
static int32_t doMinMaxScalarFunction(SScalarParam *pInput, int32_t inputNum, SScalarParam *pOutput, bool isMinFunc) {
H
Hongze Cheng 已提交
1812
  SColumnInfoData *pInputData = pInput->columnData;
1813 1814 1815
  SColumnInfoData *pOutputData = pOutput->columnData;

  int32_t type = GET_PARAM_TYPE(pInput);
1816

G
Ganlin Zhao 已提交
1817
  bool hasNull = false;
1818 1819 1820 1821 1822
  if (isMinFunc) {
    SET_TYPED_DATA_MAX(pOutputData->pData, type);
  } else {
    SET_TYPED_DATA_MIN(pOutputData->pData, type);
  }
1823 1824 1825

  for (int32_t i = 0; i < pInput->numOfRows; ++i) {
    if (colDataIsNull_s(pInputData, i)) {
G
Ganlin Zhao 已提交
1826
      hasNull = true;
1827 1828 1829
      break;
    }

H
Hongze Cheng 已提交
1830
    switch (type) {
1831 1832
      case TSDB_DATA_TYPE_BOOL:
      case TSDB_DATA_TYPE_TINYINT: {
H
Hongze Cheng 已提交
1833
        int8_t *in = (int8_t *)pInputData->pData;
1834
        int8_t *out = (int8_t *)pOutputData->pData;
H
Hongze Cheng 已提交
1835
        if ((in[i] > *out) ^ isMinFunc) {
1836 1837 1838 1839 1840
          *out = in[i];
        }
        break;
      }
      case TSDB_DATA_TYPE_SMALLINT: {
H
Hongze Cheng 已提交
1841
        int16_t *in = (int16_t *)pInputData->pData;
1842
        int16_t *out = (int16_t *)pOutputData->pData;
H
Hongze Cheng 已提交
1843
        if ((in[i] > *out) ^ isMinFunc) {
1844 1845 1846 1847 1848
          *out = in[i];
        }
        break;
      }
      case TSDB_DATA_TYPE_INT: {
H
Hongze Cheng 已提交
1849
        int32_t *in = (int32_t *)pInputData->pData;
1850
        int32_t *out = (int32_t *)pOutputData->pData;
H
Hongze Cheng 已提交
1851
        if ((in[i] > *out) ^ isMinFunc) {
1852 1853 1854 1855 1856
          *out = in[i];
        }
        break;
      }
      case TSDB_DATA_TYPE_BIGINT: {
H
Hongze Cheng 已提交
1857
        int64_t *in = (int64_t *)pInputData->pData;
1858
        int64_t *out = (int64_t *)pOutputData->pData;
H
Hongze Cheng 已提交
1859
        if ((in[i] > *out) ^ isMinFunc) {
1860 1861 1862 1863 1864
          *out = in[i];
        }
        break;
      }
      case TSDB_DATA_TYPE_UTINYINT: {
H
Hongze Cheng 已提交
1865
        uint8_t *in = (uint8_t *)pInputData->pData;
1866
        uint8_t *out = (uint8_t *)pOutputData->pData;
H
Hongze Cheng 已提交
1867
        if ((in[i] > *out) ^ isMinFunc) {
1868 1869 1870 1871 1872
          *out = in[i];
        }
        break;
      }
      case TSDB_DATA_TYPE_USMALLINT: {
H
Hongze Cheng 已提交
1873
        uint16_t *in = (uint16_t *)pInputData->pData;
1874
        uint16_t *out = (uint16_t *)pOutputData->pData;
H
Hongze Cheng 已提交
1875
        if ((in[i] > *out) ^ isMinFunc) {
1876 1877 1878 1879 1880
          *out = in[i];
        }
        break;
      }
      case TSDB_DATA_TYPE_UINT: {
H
Hongze Cheng 已提交
1881
        uint32_t *in = (uint32_t *)pInputData->pData;
1882
        uint32_t *out = (uint32_t *)pOutputData->pData;
H
Hongze Cheng 已提交
1883
        if ((in[i] > *out) ^ isMinFunc) {
1884 1885 1886 1887 1888
          *out = in[i];
        }
        break;
      }
      case TSDB_DATA_TYPE_UBIGINT: {
H
Hongze Cheng 已提交
1889
        uint64_t *in = (uint64_t *)pInputData->pData;
1890
        uint64_t *out = (uint64_t *)pOutputData->pData;
H
Hongze Cheng 已提交
1891
        if ((in[i] > *out) ^ isMinFunc) {
1892 1893 1894 1895 1896
          *out = in[i];
        }
        break;
      }
      case TSDB_DATA_TYPE_FLOAT: {
H
Hongze Cheng 已提交
1897
        float *in = (float *)pInputData->pData;
1898
        float *out = (float *)pOutputData->pData;
H
Hongze Cheng 已提交
1899
        if ((in[i] > *out) ^ isMinFunc) {
1900 1901 1902 1903 1904
          *out = in[i];
        }
        break;
      }
      case TSDB_DATA_TYPE_DOUBLE: {
H
Hongze Cheng 已提交
1905
        double *in = (double *)pInputData->pData;
1906
        double *out = (double *)pOutputData->pData;
H
Hongze Cheng 已提交
1907
        if ((in[i] > *out) ^ isMinFunc) {
1908 1909 1910 1911 1912 1913 1914
          *out = in[i];
        }
        break;
      }
    }
  }

G
Ganlin Zhao 已提交
1915 1916 1917 1918 1919
  if (hasNull) {
    colDataAppendNULL(pOutputData, 0);
  }

  pOutput->numOfRows = 1;
1920 1921
  return TSDB_CODE_SUCCESS;
}
1922 1923 1924 1925 1926 1927 1928 1929

int32_t minScalarFunction(SScalarParam *pInput, int32_t inputNum, SScalarParam *pOutput) {
  return doMinMaxScalarFunction(pInput, inputNum, pOutput, true);
}

int32_t maxScalarFunction(SScalarParam *pInput, int32_t inputNum, SScalarParam *pOutput) {
  return doMinMaxScalarFunction(pInput, inputNum, pOutput, false);
}
1930

1931
int32_t avgScalarFunction(SScalarParam *pInput, int32_t inputNum, SScalarParam *pOutput) {
H
Hongze Cheng 已提交
1932
  SColumnInfoData *pInputData = pInput->columnData;
1933 1934 1935
  SColumnInfoData *pOutputData = pOutput->columnData;

  int32_t type = GET_PARAM_TYPE(pInput);
1936
  int64_t count = 0;
H
Hongze Cheng 已提交
1937
  bool    hasNull = false;
1938 1939 1940 1941 1942 1943 1944

  for (int32_t i = 0; i < pInput->numOfRows; ++i) {
    if (colDataIsNull_s(pInputData, i)) {
      hasNull = true;
      break;
    }

H
Hongze Cheng 已提交
1945
    switch (type) {
1946
      case TSDB_DATA_TYPE_TINYINT: {
H
Hongze Cheng 已提交
1947
        int8_t  *in = (int8_t *)pInputData->pData;
1948 1949
        int64_t *out = (int64_t *)pOutputData->pData;
        *out += in[i];
1950 1951 1952 1953
        count++;
        break;
      }
      case TSDB_DATA_TYPE_SMALLINT: {
H
Hongze Cheng 已提交
1954
        int16_t *in = (int16_t *)pInputData->pData;
1955 1956
        int64_t *out = (int64_t *)pOutputData->pData;
        *out += in[i];
1957 1958 1959 1960
        count++;
        break;
      }
      case TSDB_DATA_TYPE_INT: {
H
Hongze Cheng 已提交
1961
        int32_t *in = (int32_t *)pInputData->pData;
1962 1963
        int64_t *out = (int64_t *)pOutputData->pData;
        *out += in[i];
1964 1965 1966 1967
        count++;
        break;
      }
      case TSDB_DATA_TYPE_BIGINT: {
H
Hongze Cheng 已提交
1968
        int64_t *in = (int64_t *)pInputData->pData;
1969 1970
        int64_t *out = (int64_t *)pOutputData->pData;
        *out += in[i];
1971 1972 1973 1974
        count++;
        break;
      }
      case TSDB_DATA_TYPE_UTINYINT: {
H
Hongze Cheng 已提交
1975
        uint8_t  *in = (uint8_t *)pInputData->pData;
1976 1977
        uint64_t *out = (uint64_t *)pOutputData->pData;
        *out += in[i];
1978 1979 1980 1981
        count++;
        break;
      }
      case TSDB_DATA_TYPE_USMALLINT: {
H
Hongze Cheng 已提交
1982
        uint16_t *in = (uint16_t *)pInputData->pData;
1983 1984
        uint64_t *out = (uint64_t *)pOutputData->pData;
        *out += in[i];
1985 1986 1987 1988
        count++;
        break;
      }
      case TSDB_DATA_TYPE_UINT: {
H
Hongze Cheng 已提交
1989
        uint32_t *in = (uint32_t *)pInputData->pData;
1990 1991
        uint64_t *out = (uint64_t *)pOutputData->pData;
        *out += in[i];
1992 1993 1994 1995
        count++;
        break;
      }
      case TSDB_DATA_TYPE_UBIGINT: {
H
Hongze Cheng 已提交
1996
        uint64_t *in = (uint64_t *)pInputData->pData;
1997 1998
        uint64_t *out = (uint64_t *)pOutputData->pData;
        *out += in[i];
1999 2000 2001 2002
        count++;
        break;
      }
      case TSDB_DATA_TYPE_FLOAT: {
H
Hongze Cheng 已提交
2003
        float *in = (float *)pInputData->pData;
2004 2005
        float *out = (float *)pOutputData->pData;
        *out += in[i];
2006 2007 2008 2009
        count++;
        break;
      }
      case TSDB_DATA_TYPE_DOUBLE: {
H
Hongze Cheng 已提交
2010
        double *in = (double *)pInputData->pData;
2011 2012
        double *out = (double *)pOutputData->pData;
        *out += in[i];
2013 2014 2015 2016 2017 2018
        count++;
        break;
      }
    }
  }

G
Ganlin Zhao 已提交
2019
  if (hasNull || (count == 0)) {
2020 2021 2022
    colDataAppendNULL(pOutputData, 0);
  } else {
    if (IS_SIGNED_NUMERIC_TYPE(type)) {
H
Hongze Cheng 已提交
2023
      int64_t *out = (int64_t *)pOutputData->pData;
2024
      *(double *)out = *out / (double)count;
2025
    } else if (IS_UNSIGNED_NUMERIC_TYPE(type)) {
H
Hongze Cheng 已提交
2026
      uint64_t *out = (uint64_t *)pOutputData->pData;
2027
      *(double *)out = *out / (double)count;
2028
    } else if (IS_FLOAT_TYPE(type)) {
H
Hongze Cheng 已提交
2029
      double *out = (double *)pOutputData->pData;
2030
      *(double *)out = *out / (double)count;
2031 2032 2033 2034 2035 2036 2037
    }
  }

  pOutput->numOfRows = 1;
  return TSDB_CODE_SUCCESS;
}

2038
int32_t stddevScalarFunction(SScalarParam *pInput, int32_t inputNum, SScalarParam *pOutput) {
H
Hongze Cheng 已提交
2039
  SColumnInfoData *pInputData = pInput->columnData;
2040 2041 2042
  SColumnInfoData *pOutputData = pOutput->columnData;

  int32_t type = GET_PARAM_TYPE(pInput);
H
Hongze Cheng 已提交
2043
  // int64_t count = 0, sum = 0, qSum = 0;
G
Ganlin Zhao 已提交
2044
  bool hasNull = false;
2045 2046 2047

  for (int32_t i = 0; i < pInput->numOfRows; ++i) {
    if (colDataIsNull_s(pInputData, i)) {
G
Ganlin Zhao 已提交
2048 2049
      hasNull = true;
      break;
2050 2051 2052 2053 2054 2055 2056 2057 2058 2059 2060 2061 2062 2063 2064 2065 2066 2067 2068 2069 2070 2071 2072 2073 2074 2075 2076 2077 2078 2079 2080 2081 2082 2083 2084 2085 2086 2087 2088 2089 2090 2091 2092 2093 2094 2095 2096 2097 2098 2099 2100 2101 2102 2103 2104 2105 2106 2107 2108 2109 2110 2111 2112 2113 2114 2115 2116 2117 2118 2119 2120 2121 2122 2123 2124 2125 2126
    }
#if 0
    switch(type) {
      case TSDB_DATA_TYPE_TINYINT: {
        int8_t *in  = (int8_t *)pInputData->pData;
        sum += in[i];
        qSum += in[i] * in[i];
        count++;
        break;
      }
      case TSDB_DATA_TYPE_SMALLINT: {
        int16_t *in  = (int16_t *)pInputData->pData;
        sum += in[i];
        qSum += in[i] * in[i];
        count++;
        break;
      }
      case TSDB_DATA_TYPE_INT: {
        int32_t *in  = (int32_t *)pInputData->pData;
        sum += in[i];
        qSum += in[i] * in[i];
        count++;
        break;
      }
      case TSDB_DATA_TYPE_BIGINT: {
        int64_t *in  = (int64_t *)pInputData->pData;
        sum += in[i];
        qSum += in[i] * in[i];
        count++;
        break;
      }
      case TSDB_DATA_TYPE_UTINYINT: {
        uint8_t *in  = (uint8_t *)pInputData->pData;
        sum += in[i];
        qSum += in[i] * in[i];
        count++;
        break;
      }
      case TSDB_DATA_TYPE_USMALLINT: {
        uint16_t *in  = (uint16_t *)pInputData->pData;
        sum += in[i];
        qSum += in[i] * in[i];
        count++;
        break;
      }
      case TSDB_DATA_TYPE_UINT: {
        uint32_t *in  = (uint32_t *)pInputData->pData;
        sum += in[i];
        qSum += in[i] * in[i];
        count++;
        break;
      }
      case TSDB_DATA_TYPE_UBIGINT: {
        uint64_t *in  = (uint64_t *)pInputData->pData;
        sum += in[i];
        qSum += in[i] * in[i];
        count++;
        break;
      }
      case TSDB_DATA_TYPE_FLOAT: {
        float *in  = (float *)pInputData->pData;
        sum += in[i];
        qSum += in[i] * in[i];
        count++;
        break;
      }
      case TSDB_DATA_TYPE_DOUBLE: {
        double *in  = (double *)pInputData->pData;
        sum += in[i];
        qSum += in[i] * in[i];
        count++;
        break;
      }
    }
#endif
  }

H
Hongze Cheng 已提交
2127
  double *out = (double *)pOutputData->pData;
G
Ganlin Zhao 已提交
2128
  if (hasNull) {
2129 2130 2131 2132 2133 2134 2135 2136 2137 2138 2139 2140 2141 2142 2143 2144 2145 2146
    colDataAppendNULL(pOutputData, 0);
  } else {
    *out = 0;
#if 0
    double avg = 0;
    if (IS_SIGNED_NUMERIC_TYPE(type)) {
      avg = (int64_t)sum / (double)count;
      *out =  sqrt(fabs((int64_t)qSum / ((double)count) - avg * avg));
    } else if (IS_UNSIGNED_NUMERIC_TYPE(type)) {
      avg = (uint64_t)sum / (double)count;
      *out =  sqrt(fabs((uint64_t)qSum / ((double)count) - avg * avg));
    } else if (IS_FLOAT_TYPE(type)) {
      avg = (double)sum / (double)count;
      *out =  sqrt(fabs((double)qSum / ((double)count) - avg * avg));
    }
#endif
  }

G
Ganlin Zhao 已提交
2147
  pOutput->numOfRows = 1;
2148 2149
  return TSDB_CODE_SUCCESS;
}
2150

2151 2152 2153 2154 2155 2156 2157 2158 2159 2160
#define LEASTSQR_CAL(p, x, y, index, step) \
  do {                                     \
    (p)[0][0] += (double)(x) * (x);        \
    (p)[0][1] += (double)(x);              \
    (p)[0][2] += (double)(x) * (y)[index]; \
    (p)[1][2] += (y)[index];               \
    (x) += step;                           \
  } while (0)

int32_t leastSQRScalarFunction(SScalarParam *pInput, int32_t inputNum, SScalarParam *pOutput) {
H
Hongze Cheng 已提交
2161
  SColumnInfoData *pInputData = pInput->columnData;
2162 2163 2164 2165 2166 2167 2168 2169 2170 2171
  SColumnInfoData *pOutputData = pOutput->columnData;

  double startVal, stepVal;
  double matrix[2][3] = {0};
  GET_TYPED_DATA(startVal, double, GET_PARAM_TYPE(&pInput[1]), pInput[1].columnData->pData);
  GET_TYPED_DATA(stepVal, double, GET_PARAM_TYPE(&pInput[2]), pInput[2].columnData->pData);

  int32_t type = GET_PARAM_TYPE(pInput);
  int64_t count = 0;

H
Hongze Cheng 已提交
2172
  switch (type) {
2173
    case TSDB_DATA_TYPE_TINYINT: {
H
Hongze Cheng 已提交
2174
      int8_t *in = (int8_t *)pInputData->pData;
2175 2176 2177 2178 2179 2180 2181 2182 2183 2184 2185
      for (int32_t i = 0; i < pInput->numOfRows; ++i) {
        if (colDataIsNull_s(pInputData, i)) {
          continue;
        }

        count++;
        LEASTSQR_CAL(matrix, startVal, in, i, stepVal);
      }
      break;
    }
    case TSDB_DATA_TYPE_SMALLINT: {
H
Hongze Cheng 已提交
2186
      int16_t *in = (int16_t *)pInputData->pData;
2187 2188 2189 2190 2191 2192 2193 2194 2195 2196 2197
      for (int32_t i = 0; i < pInput->numOfRows; ++i) {
        if (colDataIsNull_s(pInputData, i)) {
          continue;
        }

        count++;
        LEASTSQR_CAL(matrix, startVal, in, i, stepVal);
      }
      break;
    }
    case TSDB_DATA_TYPE_INT: {
H
Hongze Cheng 已提交
2198
      int32_t *in = (int32_t *)pInputData->pData;
2199 2200 2201 2202 2203 2204 2205 2206 2207 2208 2209
      for (int32_t i = 0; i < pInput->numOfRows; ++i) {
        if (colDataIsNull_s(pInputData, i)) {
          continue;
        }

        count++;
        LEASTSQR_CAL(matrix, startVal, in, i, stepVal);
      }
      break;
    }
    case TSDB_DATA_TYPE_BIGINT: {
H
Hongze Cheng 已提交
2210
      int64_t *in = (int64_t *)pInputData->pData;
2211 2212 2213 2214 2215 2216 2217 2218 2219 2220 2221
      for (int32_t i = 0; i < pInput->numOfRows; ++i) {
        if (colDataIsNull_s(pInputData, i)) {
          continue;
        }

        count++;
        LEASTSQR_CAL(matrix, startVal, in, i, stepVal);
      }
      break;
    }
    case TSDB_DATA_TYPE_UTINYINT: {
H
Hongze Cheng 已提交
2222
      uint8_t *in = (uint8_t *)pInputData->pData;
2223 2224 2225 2226 2227 2228 2229 2230 2231 2232 2233
      for (int32_t i = 0; i < pInput->numOfRows; ++i) {
        if (colDataIsNull_s(pInputData, i)) {
          continue;
        }

        count++;
        LEASTSQR_CAL(matrix, startVal, in, i, stepVal);
      }
      break;
    }
    case TSDB_DATA_TYPE_USMALLINT: {
H
Hongze Cheng 已提交
2234
      uint16_t *in = (uint16_t *)pInputData->pData;
2235 2236 2237 2238 2239 2240 2241 2242 2243 2244 2245
      for (int32_t i = 0; i < pInput->numOfRows; ++i) {
        if (colDataIsNull_s(pInputData, i)) {
          continue;
        }

        count++;
        LEASTSQR_CAL(matrix, startVal, in, i, stepVal);
      }
      break;
    }
    case TSDB_DATA_TYPE_UINT: {
H
Hongze Cheng 已提交
2246
      uint32_t *in = (uint32_t *)pInputData->pData;
2247 2248 2249 2250 2251 2252 2253 2254 2255 2256 2257
      for (int32_t i = 0; i < pInput->numOfRows; ++i) {
        if (colDataIsNull_s(pInputData, i)) {
          continue;
        }

        count++;
        LEASTSQR_CAL(matrix, startVal, in, i, stepVal);
      }
      break;
    }
    case TSDB_DATA_TYPE_UBIGINT: {
H
Hongze Cheng 已提交
2258
      uint64_t *in = (uint64_t *)pInputData->pData;
2259 2260 2261 2262 2263 2264 2265 2266 2267 2268 2269
      for (int32_t i = 0; i < pInput->numOfRows; ++i) {
        if (colDataIsNull_s(pInputData, i)) {
          continue;
        }

        count++;
        LEASTSQR_CAL(matrix, startVal, in, i, stepVal);
      }
      break;
    }
    case TSDB_DATA_TYPE_FLOAT: {
H
Hongze Cheng 已提交
2270
      float *in = (float *)pInputData->pData;
2271 2272 2273 2274 2275 2276 2277 2278 2279 2280 2281
      for (int32_t i = 0; i < pInput->numOfRows; ++i) {
        if (colDataIsNull_s(pInputData, i)) {
          continue;
        }

        count++;
        LEASTSQR_CAL(matrix, startVal, in, i, stepVal);
      }
      break;
    }
    case TSDB_DATA_TYPE_DOUBLE: {
H
Hongze Cheng 已提交
2282
      double *in = (double *)pInputData->pData;
2283 2284 2285 2286 2287 2288 2289 2290 2291 2292 2293 2294 2295 2296 2297 2298 2299 2300 2301 2302 2303 2304 2305 2306 2307 2308
      for (int32_t i = 0; i < pInput->numOfRows; ++i) {
        if (colDataIsNull_s(pInputData, i)) {
          continue;
        }

        count++;
        LEASTSQR_CAL(matrix, startVal, in, i, stepVal);
      }
      break;
    }
  }

  if (count == 0) {
    colDataAppendNULL(pOutputData, 0);
  } else {
    matrix[1][1] = (double)count;
    matrix[1][0] = matrix[0][1];

    double matrix00 = matrix[0][0] - matrix[1][0] * (matrix[0][1] / matrix[1][1]);
    double matrix02 = matrix[0][2] - matrix[1][2] * (matrix[0][1] / matrix[1][1]);
    double matrix12 = matrix[1][2] - matrix02 * (matrix[1][0] / matrix00);
    matrix02 /= matrix00;

    matrix12 /= matrix[1][1];

    char   buf[64] = {0};
H
Hongze Cheng 已提交
2309 2310
    size_t len = snprintf(varDataVal(buf), sizeof(buf) - VARSTR_HEADER_SIZE, "{slop:%.6lf, intercept:%.6lf}", matrix02,
                          matrix12);
2311 2312 2313 2314 2315 2316 2317
    varDataSetLen(buf, len);
    colDataAppend(pOutputData, 0, buf, false);
  }

  pOutput->numOfRows = 1;
  return TSDB_CODE_SUCCESS;
}
2318 2319

int32_t percentileScalarFunction(SScalarParam *pInput, int32_t inputNum, SScalarParam *pOutput) {
H
Hongze Cheng 已提交
2320
  SColumnInfoData *pInputData = pInput->columnData;
2321 2322 2323 2324 2325
  SColumnInfoData *pOutputData = pOutput->columnData;

  int32_t type = GET_PARAM_TYPE(pInput);

  double val;
H
Hongze Cheng 已提交
2326
  bool   hasNull = false;
2327 2328 2329 2330 2331 2332 2333 2334 2335 2336 2337 2338 2339 2340 2341 2342 2343 2344 2345 2346 2347 2348
  for (int32_t i = 0; i < pInput->numOfRows; ++i) {
    if (colDataIsNull_s(pInputData, i)) {
      hasNull = true;
      break;
    }
    char *in = pInputData->pData;
    GET_TYPED_DATA(val, double, type, in);
  }

  if (hasNull) {
    colDataAppendNULL(pOutputData, 0);
  } else {
    colDataAppend(pOutputData, 0, (char *)&val, false);
  }

  pOutput->numOfRows = 1;
  return TSDB_CODE_SUCCESS;
}

int32_t apercentileScalarFunction(SScalarParam *pInput, int32_t inputNum, SScalarParam *pOutput) {
  return percentileScalarFunction(pInput, inputNum, pOutput);
}
2349 2350

int32_t spreadScalarFunction(SScalarParam *pInput, int32_t inputNum, SScalarParam *pOutput) {
H
Hongze Cheng 已提交
2351
  SColumnInfoData *pInputData = pInput->columnData;
2352 2353 2354 2355 2356 2357 2358 2359 2360 2361 2362 2363 2364 2365 2366 2367 2368 2369 2370 2371 2372 2373 2374 2375 2376 2377 2378 2379 2380 2381 2382 2383 2384 2385 2386 2387 2388 2389 2390
  SColumnInfoData *pOutputData = pOutput->columnData;

  int32_t type = GET_PARAM_TYPE(pInput);

  double min, max;
  SET_DOUBLE_VAL(&min, DBL_MAX);
  SET_DOUBLE_VAL(&max, -DBL_MAX);

  bool hasNull = false;
  for (int32_t i = 0; i < pInput->numOfRows; ++i) {
    if (colDataIsNull_s(pInputData, i)) {
      hasNull = true;
      break;
    }

    char *in = pInputData->pData;

    double val = 0;
    GET_TYPED_DATA(val, double, type, in);

    if (val < GET_DOUBLE_VAL(&min)) {
      SET_DOUBLE_VAL(&min, val);
    }

    if (val > GET_DOUBLE_VAL(&max)) {
      SET_DOUBLE_VAL(&max, val);
    }
  }

  if (hasNull) {
    colDataAppendNULL(pOutputData, 0);
  } else {
    double result = max - min;
    colDataAppend(pOutputData, 0, (char *)&result, false);
  }

  pOutput->numOfRows = 1;
  return TSDB_CODE_SUCCESS;
}
2391

G
Ganlin Zhao 已提交
2392
int32_t nonCalcScalarFunction(SScalarParam *pInput, int32_t inputNum, SScalarParam *pOutput) {
H
Hongze Cheng 已提交
2393
  SColumnInfoData *pInputData = pInput->columnData;
2394 2395 2396
  SColumnInfoData *pOutputData = pOutput->columnData;

  int32_t type = GET_PARAM_TYPE(pInput);
H
Hongze Cheng 已提交
2397
  bool    hasNull = false;
2398 2399 2400 2401 2402 2403 2404 2405

  for (int32_t i = 0; i < pInput->numOfRows; ++i) {
    if (colDataIsNull_s(pInputData, i)) {
      hasNull = true;
      break;
    }
  }

H
Hongze Cheng 已提交
2406
  double *out = (double *)pOutputData->pData;
2407 2408 2409 2410 2411 2412 2413 2414 2415
  if (hasNull) {
    colDataAppendNULL(pOutputData, 0);
  } else {
    *out = 0;
  }

  pOutput->numOfRows = 1;
  return TSDB_CODE_SUCCESS;
}
G
Ganlin Zhao 已提交
2416 2417 2418 2419 2420 2421 2422 2423

int32_t derivativeScalarFunction(SScalarParam *pInput, int32_t inputNum, SScalarParam *pOutput) {
  return nonCalcScalarFunction(pInput, inputNum, pOutput);
}

int32_t irateScalarFunction(SScalarParam *pInput, int32_t inputNum, SScalarParam *pOutput) {
  return nonCalcScalarFunction(pInput, inputNum, pOutput);
}
2424

2425 2426 2427 2428
int32_t diffScalarFunction(SScalarParam *pInput, int32_t inputNum, SScalarParam *pOutput) {
  return nonCalcScalarFunction(pInput, inputNum, pOutput);
}

2429 2430 2431
int32_t twaScalarFunction(SScalarParam *pInput, int32_t inputNum, SScalarParam *pOutput) {
  return avgScalarFunction(pInput, inputNum, pOutput);
}
2432 2433 2434 2435

int32_t mavgScalarFunction(SScalarParam *pInput, int32_t inputNum, SScalarParam *pOutput) {
  return avgScalarFunction(pInput, inputNum, pOutput);
}
2436 2437 2438 2439

int32_t hllScalarFunction(SScalarParam *pInput, int32_t inputNum, SScalarParam *pOutput) {
  return countScalarFunction(pInput, inputNum, pOutput);
}
2440 2441 2442 2443

int32_t csumScalarFunction(SScalarParam *pInput, int32_t inputNum, SScalarParam *pOutput) {
  return sumScalarFunction(pInput, inputNum, pOutput);
}
2444 2445 2446 2447 2448 2449 2450 2451 2452 2453 2454 2455 2456 2457 2458 2459 2460 2461 2462 2463 2464 2465 2466 2467 2468 2469 2470 2471 2472 2473 2474 2475 2476 2477 2478 2479 2480 2481 2482 2483 2484

typedef enum {
  STATE_OPER_INVALID = 0,
  STATE_OPER_LT,
  STATE_OPER_GT,
  STATE_OPER_LE,
  STATE_OPER_GE,
  STATE_OPER_NE,
  STATE_OPER_EQ,
} EStateOperType;

#define STATE_COMP(_op, _lval, _rval, _rtype) STATE_COMP_IMPL(_op, _lval, GET_STATE_VAL(_rval, _rtype))

#define GET_STATE_VAL(_val, _type) ((_type == TSDB_DATA_TYPE_BIGINT) ? (*(int64_t *)_val) : (*(double *)_val))

#define STATE_COMP_IMPL(_op, _lval, _rval) \
  do {                                     \
    switch (_op) {                         \
      case STATE_OPER_LT:                  \
        return ((_lval) < (_rval));        \
        break;                             \
      case STATE_OPER_GT:                  \
        return ((_lval) > (_rval));        \
        break;                             \
      case STATE_OPER_LE:                  \
        return ((_lval) <= (_rval));       \
        break;                             \
      case STATE_OPER_GE:                  \
        return ((_lval) >= (_rval));       \
        break;                             \
      case STATE_OPER_NE:                  \
        return ((_lval) != (_rval));       \
        break;                             \
      case STATE_OPER_EQ:                  \
        return ((_lval) == (_rval));       \
        break;                             \
      default:                             \
        break;                             \
    }                                      \
  } while (0)

H
Hongze Cheng 已提交
2485
static int8_t getStateOpType(char *opStr) {
2486
  int8_t opType;
2487
  if (strncasecmp(opStr, "LT", 2) == 0) {
2488
    opType = STATE_OPER_LT;
2489
  } else if (strncasecmp(opStr, "GT", 2) == 0) {
2490
    opType = STATE_OPER_GT;
2491
  } else if (strncasecmp(opStr, "LE", 2) == 0) {
2492
    opType = STATE_OPER_LE;
2493
  } else if (strncasecmp(opStr, "GE", 2) == 0) {
2494
    opType = STATE_OPER_GE;
2495
  } else if (strncasecmp(opStr, "NE", 2) == 0) {
2496
    opType = STATE_OPER_NE;
2497
  } else if (strncasecmp(opStr, "EQ", 2) == 0) {
2498 2499 2500 2501 2502 2503 2504 2505
    opType = STATE_OPER_EQ;
  } else {
    opType = STATE_OPER_INVALID;
  }

  return opType;
}

H
Hongze Cheng 已提交
2506 2507 2508
static bool checkStateOp(int8_t op, SColumnInfoData *pCol, int32_t index, SScalarParam *pCondParam) {
  char   *data = colDataGetData(pCol, index);
  char   *param = pCondParam->columnData->pData;
2509 2510 2511
  int32_t paramType = GET_PARAM_TYPE(pCondParam);
  switch (pCol->info.type) {
    case TSDB_DATA_TYPE_TINYINT: {
H
Hongze Cheng 已提交
2512
      int8_t v = *(int8_t *)data;
2513 2514 2515 2516
      STATE_COMP(op, v, param, paramType);
      break;
    }
    case TSDB_DATA_TYPE_UTINYINT: {
H
Hongze Cheng 已提交
2517
      uint8_t v = *(uint8_t *)data;
2518 2519 2520 2521
      STATE_COMP(op, v, param, paramType);
      break;
    }
    case TSDB_DATA_TYPE_SMALLINT: {
H
Hongze Cheng 已提交
2522
      int16_t v = *(int16_t *)data;
2523 2524 2525 2526
      STATE_COMP(op, v, param, paramType);
      break;
    }
    case TSDB_DATA_TYPE_USMALLINT: {
H
Hongze Cheng 已提交
2527
      uint16_t v = *(uint16_t *)data;
2528 2529 2530 2531
      STATE_COMP(op, v, param, paramType);
      break;
    }
    case TSDB_DATA_TYPE_INT: {
H
Hongze Cheng 已提交
2532
      int32_t v = *(int32_t *)data;
2533 2534 2535 2536
      STATE_COMP(op, v, param, paramType);
      break;
    }
    case TSDB_DATA_TYPE_UINT: {
H
Hongze Cheng 已提交
2537
      uint32_t v = *(uint32_t *)data;
2538 2539 2540 2541
      STATE_COMP(op, v, param, paramType);
      break;
    }
    case TSDB_DATA_TYPE_BIGINT: {
H
Hongze Cheng 已提交
2542
      int64_t v = *(int64_t *)data;
2543 2544 2545 2546
      STATE_COMP(op, v, param, paramType);
      break;
    }
    case TSDB_DATA_TYPE_UBIGINT: {
H
Hongze Cheng 已提交
2547
      uint64_t v = *(uint64_t *)data;
2548 2549 2550 2551
      STATE_COMP(op, v, param, paramType);
      break;
    }
    case TSDB_DATA_TYPE_FLOAT: {
H
Hongze Cheng 已提交
2552
      float v = *(float *)data;
2553 2554 2555 2556
      STATE_COMP(op, v, param, paramType);
      break;
    }
    case TSDB_DATA_TYPE_DOUBLE: {
H
Hongze Cheng 已提交
2557
      double v = *(double *)data;
2558 2559 2560 2561 2562 2563 2564 2565 2566 2567 2568
      STATE_COMP(op, v, param, paramType);
      break;
    }
    default: {
      ASSERT(0);
    }
  }
  return false;
}

int32_t stateCountScalarFunction(SScalarParam *pInput, int32_t inputNum, SScalarParam *pOutput) {
H
Hongze Cheng 已提交
2569
  SColumnInfoData *pInputData = pInput->columnData;
2570 2571
  SColumnInfoData *pOutputData = pOutput->columnData;

H
Hongze Cheng 已提交
2572
  int8_t  op = getStateOpType(varDataVal(pInput[1].columnData->pData));
2573 2574 2575 2576 2577 2578 2579 2580
  int64_t count = 0;

  for (int32_t i = 0; i < pInput->numOfRows; ++i) {
    if (colDataIsNull_s(pInputData, i)) {
      colDataAppendNULL(pOutputData, i);
      continue;
    }

H
Hongze Cheng 已提交
2581
    bool    ret = checkStateOp(op, pInputData, i, &pInput[2]);
2582 2583 2584 2585 2586 2587
    int64_t out = -1;
    if (ret) {
      out = ++count;
    } else {
      count = 0;
    }
H
Hongze Cheng 已提交
2588
    colDataAppend(pOutputData, i, (char *)&out, false);
2589 2590 2591 2592 2593
  }

  pOutput->numOfRows = pInput->numOfRows;
  return TSDB_CODE_SUCCESS;
}
2594 2595

int32_t stateDurationScalarFunction(SScalarParam *pInput, int32_t inputNum, SScalarParam *pOutput) {
H
Hongze Cheng 已提交
2596
  SColumnInfoData *pInputData = pInput->columnData;
2597 2598 2599 2600 2601 2602 2603 2604 2605 2606
  SColumnInfoData *pOutputData = pOutput->columnData;

  int8_t op = getStateOpType(varDataVal(pInput[1].columnData->pData));

  for (int32_t i = 0; i < pInput->numOfRows; ++i) {
    if (colDataIsNull_s(pInputData, i)) {
      colDataAppendNULL(pOutputData, i);
      continue;
    }

H
Hongze Cheng 已提交
2607
    bool    ret = checkStateOp(op, pInputData, i, &pInput[2]);
2608 2609 2610 2611
    int64_t out = -1;
    if (ret) {
      out = 0;
    }
H
Hongze Cheng 已提交
2612
    colDataAppend(pOutputData, i, (char *)&out, false);
2613 2614 2615 2616 2617
  }

  pOutput->numOfRows = pInput->numOfRows;
  return TSDB_CODE_SUCCESS;
}
2618 2619 2620

typedef enum { UNKNOWN_BIN = 0, USER_INPUT_BIN, LINEAR_BIN, LOG_BIN } EHistoBinType;

H
Hongze Cheng 已提交
2621
static int8_t getHistogramBinType(char *binTypeStr) {
2622 2623 2624 2625 2626 2627 2628 2629 2630 2631 2632 2633 2634 2635 2636 2637 2638 2639 2640 2641 2642
  int8_t binType;
  if (strcasecmp(binTypeStr, "user_input") == 0) {
    binType = USER_INPUT_BIN;
  } else if (strcasecmp(binTypeStr, "linear_bin") == 0) {
    binType = LINEAR_BIN;
  } else if (strcasecmp(binTypeStr, "log_bin") == 0) {
    binType = LOG_BIN;
  } else {
    binType = UNKNOWN_BIN;
  }

  return binType;
}

typedef struct SHistoFuncBin {
  double  lower;
  double  upper;
  int64_t count;
  double  percentage;
} SHistoFuncBin;

H
Hongze Cheng 已提交
2643 2644 2645
static bool getHistogramBinDesc(SHistoFuncBin **bins, int32_t *binNum, char *binDescStr, int8_t binType,
                                bool normalized) {
  cJSON  *binDesc = cJSON_Parse(binDescStr);
2646
  int32_t numOfBins;
H
Hongze Cheng 已提交
2647
  double *intervals;
2648 2649 2650 2651
  if (cJSON_IsObject(binDesc)) { /* linaer/log bins */
    int32_t numOfParams = cJSON_GetArraySize(binDesc);
    int32_t startIndex;
    if (numOfParams != 4) {
2652
      cJSON_Delete(binDesc);
2653 2654 2655
      return false;
    }

H
Hongze Cheng 已提交
2656 2657 2658 2659 2660
    cJSON *start = cJSON_GetObjectItem(binDesc, "start");
    cJSON *factor = cJSON_GetObjectItem(binDesc, "factor");
    cJSON *width = cJSON_GetObjectItem(binDesc, "width");
    cJSON *count = cJSON_GetObjectItem(binDesc, "count");
    cJSON *infinity = cJSON_GetObjectItem(binDesc, "infinity");
2661 2662

    if (!cJSON_IsNumber(start) || !cJSON_IsNumber(count) || !cJSON_IsBool(infinity)) {
2663
      cJSON_Delete(binDesc);
2664 2665 2666 2667
      return false;
    }

    if (count->valueint <= 0 || count->valueint > 1000) {  // limit count to 1000
2668
      cJSON_Delete(binDesc);
2669 2670 2671 2672 2673
      return false;
    }

    if (isinf(start->valuedouble) || (width != NULL && isinf(width->valuedouble)) ||
        (factor != NULL && isinf(factor->valuedouble)) || (count != NULL && isinf(count->valuedouble))) {
2674
      cJSON_Delete(binDesc);
2675 2676 2677 2678 2679 2680 2681 2682 2683 2684 2685 2686 2687 2688 2689 2690 2691
      return false;
    }

    int32_t counter = (int32_t)count->valueint;
    if (infinity->valueint == false) {
      startIndex = 0;
      numOfBins = counter + 1;
    } else {
      startIndex = 1;
      numOfBins = counter + 3;
    }

    intervals = taosMemoryCalloc(numOfBins, sizeof(double));
    if (cJSON_IsNumber(width) && factor == NULL && binType == LINEAR_BIN) {
      // linear bin process
      if (width->valuedouble == 0) {
        taosMemoryFree(intervals);
2692
        cJSON_Delete(binDesc);
2693 2694 2695 2696 2697 2698
        return false;
      }
      for (int i = 0; i < counter + 1; ++i) {
        intervals[startIndex] = start->valuedouble + i * width->valuedouble;
        if (isinf(intervals[startIndex])) {
          taosMemoryFree(intervals);
2699
          cJSON_Delete(binDesc);
2700 2701 2702 2703 2704 2705 2706 2707
          return false;
        }
        startIndex++;
      }
    } else if (cJSON_IsNumber(factor) && width == NULL && binType == LOG_BIN) {
      // log bin process
      if (start->valuedouble == 0) {
        taosMemoryFree(intervals);
2708
        cJSON_Delete(binDesc);
2709 2710 2711 2712
        return false;
      }
      if (factor->valuedouble < 0 || factor->valuedouble == 0 || factor->valuedouble == 1) {
        taosMemoryFree(intervals);
2713
        cJSON_Delete(binDesc);
2714 2715 2716 2717 2718 2719
        return false;
      }
      for (int i = 0; i < counter + 1; ++i) {
        intervals[startIndex] = start->valuedouble * pow(factor->valuedouble, i * 1.0);
        if (isinf(intervals[startIndex])) {
          taosMemoryFree(intervals);
2720
          cJSON_Delete(binDesc);
2721 2722 2723 2724 2725 2726
          return false;
        }
        startIndex++;
      }
    } else {
      taosMemoryFree(intervals);
2727
      cJSON_Delete(binDesc);
2728 2729 2730 2731 2732 2733 2734 2735 2736 2737 2738 2739 2740 2741
      return false;
    }

    if (infinity->valueint == true) {
      intervals[0] = -INFINITY;
      intervals[numOfBins - 1] = INFINITY;
      // in case of desc bin orders, -inf/inf should be swapped
      ASSERT(numOfBins >= 4);
      if (intervals[1] > intervals[numOfBins - 2]) {
        TSWAP(intervals[0], intervals[numOfBins - 1]);
      }
    }
  } else if (cJSON_IsArray(binDesc)) { /* user input bins */
    if (binType != USER_INPUT_BIN) {
2742
      cJSON_Delete(binDesc);
2743 2744 2745 2746
      return false;
    }
    numOfBins = cJSON_GetArraySize(binDesc);
    intervals = taosMemoryCalloc(numOfBins, sizeof(double));
H
Hongze Cheng 已提交
2747
    cJSON *bin = binDesc->child;
2748 2749
    if (bin == NULL) {
      taosMemoryFree(intervals);
2750
      cJSON_Delete(binDesc);
2751 2752 2753 2754 2755 2756 2757
      return false;
    }
    int i = 0;
    while (bin) {
      intervals[i] = bin->valuedouble;
      if (!cJSON_IsNumber(bin)) {
        taosMemoryFree(intervals);
2758
        cJSON_Delete(binDesc);
2759 2760 2761 2762
        return false;
      }
      if (i != 0 && intervals[i] <= intervals[i - 1]) {
        taosMemoryFree(intervals);
2763
        cJSON_Delete(binDesc);
2764 2765 2766 2767 2768 2769
        return false;
      }
      bin = bin->next;
      i++;
    }
  } else {
2770
    cJSON_Delete(binDesc);
2771 2772 2773 2774 2775 2776 2777 2778 2779 2780 2781 2782
    return false;
  }

  *binNum = numOfBins - 1;
  *bins = taosMemoryCalloc(numOfBins, sizeof(SHistoFuncBin));
  for (int32_t i = 0; i < *binNum; ++i) {
    (*bins)[i].lower = intervals[i] < intervals[i + 1] ? intervals[i] : intervals[i + 1];
    (*bins)[i].upper = intervals[i + 1] > intervals[i] ? intervals[i + 1] : intervals[i];
    (*bins)[i].count = 0;
  }

  taosMemoryFree(intervals);
2783 2784
  cJSON_Delete(binDesc);

2785 2786 2787 2788
  return true;
}

int32_t histogramScalarFunction(SScalarParam *pInput, int32_t inputNum, SScalarParam *pOutput) {
H
Hongze Cheng 已提交
2789
  SColumnInfoData *pInputData = pInput->columnData;
2790 2791 2792
  SColumnInfoData *pOutputData = pOutput->columnData;

  SHistoFuncBin *bins;
H
Hongze Cheng 已提交
2793 2794
  int32_t        numOfBins = 0;
  int32_t        totalCount = 0;
2795

2796 2797 2798 2799 2800
  char *binTypeStr = strndup(varDataVal(pInput[1].columnData->pData), varDataLen(pInput[1].columnData->pData));
  int8_t binType = getHistogramBinType(binTypeStr);
  taosMemoryFree(binTypeStr);

  char   *binDesc = strndup(varDataVal(pInput[2].columnData->pData), varDataLen(pInput[2].columnData->pData));
2801 2802 2803 2804
  int64_t normalized = *(int64_t *)(pInput[3].columnData->pData);

  int32_t type = GET_PARAM_TYPE(pInput);
  if (!getHistogramBinDesc(&bins, &numOfBins, binDesc, binType, (bool)normalized)) {
2805
    taosMemoryFree(binDesc);
2806 2807
    return TSDB_CODE_FAILED;
  }
2808
  taosMemoryFree(binDesc);
2809 2810 2811 2812 2813 2814

  for (int32_t i = 0; i < pInput->numOfRows; ++i) {
    if (colDataIsNull_s(pInputData, i)) {
      continue;
    }

H
Hongze Cheng 已提交
2815
    char  *data = colDataGetData(pInputData, i);
2816 2817 2818 2819 2820 2821 2822 2823 2824 2825 2826 2827 2828 2829 2830 2831 2832 2833 2834 2835 2836 2837
    double v;
    GET_TYPED_DATA(v, double, type, data);

    for (int32_t k = 0; k < numOfBins; ++k) {
      if (v > bins[k].lower && v <= bins[k].upper) {
        bins[k].count++;
        totalCount++;
        break;
      }
    }
  }

  if (normalized) {
    for (int32_t k = 0; k < numOfBins; ++k) {
      if (totalCount != 0) {
        bins[k].percentage = bins[k].count / (double)totalCount;
      } else {
        bins[k].percentage = 0;
      }
    }
  }

2838 2839
  colInfoDataEnsureCapacity(pOutputData, numOfBins, false);

2840 2841 2842 2843
  for (int32_t k = 0; k < numOfBins; ++k) {
    int32_t len;
    char    buf[512] = {0};
    if (!normalized) {
H
Hongze Cheng 已提交
2844 2845
      len = sprintf(varDataVal(buf), "{\"lower_bin\":%g, \"upper_bin\":%g, \"count\":%" PRId64 "}", bins[k].lower,
                    bins[k].upper, bins[k].count);
2846
    } else {
H
Hongze Cheng 已提交
2847 2848
      len = sprintf(varDataVal(buf), "{\"lower_bin\":%g, \"upper_bin\":%g, \"count\":%lf}", bins[k].lower,
                    bins[k].upper, bins[k].percentage);
2849 2850 2851 2852 2853 2854 2855 2856 2857
    }
    varDataSetLen(buf, len);
    colDataAppend(pOutputData, k, buf, false);
  }

  taosMemoryFree(bins);
  pOutput->numOfRows = numOfBins;
  return TSDB_CODE_SUCCESS;
}
2858 2859

int32_t selectScalarFunction(SScalarParam *pInput, int32_t inputNum, SScalarParam *pOutput) {
H
Hongze Cheng 已提交
2860
  SColumnInfoData *pInputData = pInput->columnData;
2861 2862 2863 2864 2865 2866 2867 2868 2869 2870
  SColumnInfoData *pOutputData = pOutput->columnData;

  int32_t type = GET_PARAM_TYPE(pInput);

  for (int32_t i = 0; i < pInput->numOfRows; ++i) {
    if (colDataIsNull_s(pInputData, i)) {
      colDataAppendNULL(pOutputData, 0);
      continue;
    }

H
Hongze Cheng 已提交
2871
    char *data = colDataGetData(pInputData, i);
2872 2873 2874 2875 2876 2877 2878
    colDataAppend(pOutputData, i, data, false);
  }

  pOutput->numOfRows = 1;
  return TSDB_CODE_SUCCESS;
}

G
Ganlin Zhao 已提交
2879
int32_t topBotScalarFunction(SScalarParam *pInput, int32_t inputNum, SScalarParam *pOutput) {
2880 2881
  return selectScalarFunction(pInput, inputNum, pOutput);
}
2882 2883 2884 2885

int32_t firstLastScalarFunction(SScalarParam *pInput, int32_t inputNum, SScalarParam *pOutput) {
  return selectScalarFunction(pInput, inputNum, pOutput);
}
2886 2887 2888 2889 2890 2891 2892 2893 2894 2895 2896 2897 2898 2899 2900 2901

int32_t sampleScalarFunction(SScalarParam *pInput, int32_t inputNum, SScalarParam *pOutput) {
  return selectScalarFunction(pInput, inputNum, pOutput);
}

int32_t tailScalarFunction(SScalarParam *pInput, int32_t inputNum, SScalarParam *pOutput) {
  return selectScalarFunction(pInput, inputNum, pOutput);
}

int32_t uniqueScalarFunction(SScalarParam *pInput, int32_t inputNum, SScalarParam *pOutput) {
  return selectScalarFunction(pInput, inputNum, pOutput);
}

int32_t modeScalarFunction(SScalarParam *pInput, int32_t inputNum, SScalarParam *pOutput) {
  return selectScalarFunction(pInput, inputNum, pOutput);
}