tdataformat.h 24.5 KB
Newer Older
H
more  
hzcheng 已提交
1 2 3 4 5 6 7 8 9 10 11 12 13 14
/*
 * Copyright (c) 2019 TAOS Data, Inc. <jhtao@taosdata.com>
 *
 * This program is free software: you can use, redistribute, and/or modify
 * it under the terms of the GNU Affero General Public License, version 3
 * or later ("AGPL"), as published by the Free Software Foundation.
 *
 * This program is distributed in the hope that it will be useful, but WITHOUT
 * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
 * FITNESS FOR A PARTICULAR PURPOSE.
 *
 * You should have received a copy of the GNU Affero General Public License
 * along with this program. If not, see <http://www.gnu.org/licenses/>.
 */
H
hzcheng 已提交
15
#ifndef _TD_DATA_FORMAT_H_
H
more  
Hongze Cheng 已提交
16 17
#define _TD_DATA_FORMAT_H_

S
TD-4088  
Shengliang Guan 已提交
18
#include "os.h"
H
Hongze Cheng 已提交
19
#include "talgo.h"
H
Haojun Liao 已提交
20
#include "ttype.h"
H
TD-166  
hzcheng 已提交
21
#include "tutil.h"
H
hzcheng 已提交
22

H
more  
hzcheng 已提交
23 24 25
#ifdef __cplusplus
extern "C" {
#endif
H
hzcheng 已提交
26

C
Cary Xu 已提交
27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53
#pragma pack(push, 1)
typedef struct {
  VarDataLenT len;
  uint8_t     data;
} SBinaryNullT;

typedef struct {
  VarDataLenT len;
  uint32_t    data;
} SNCharNullT;
#pragma pack(pop)

extern const uint8_t      BoolNull;
extern const uint8_t      TinyintNull;
extern const uint16_t     SmallintNull;
extern const uint32_t     IntNull;
extern const uint64_t     BigintNull;
extern const uint64_t     TimestampNull;
extern const uint8_t      UTinyintNull;
extern const uint16_t     USmallintNull;
extern const uint32_t     UIntNull;
extern const uint64_t     UBigintNull;
extern const uint32_t     FloatNull;
extern const uint64_t     DoubleNull;
extern const SBinaryNullT BinaryNull;
extern const SNCharNullT  NcharNull;

54 55 56 57 58
#define STR_TO_VARSTR(x, str)                     \
  do {                                            \
    VarDataLenT __len = (VarDataLenT)strlen(str); \
    *(VarDataLenT *)(x) = __len;                  \
    memcpy(varDataVal(x), (str), __len);          \
H
Hongze Cheng 已提交
59 60
  } while (0);

61 62
#define STR_WITH_MAXSIZE_TO_VARSTR(x, str, _maxs)                         \
  do {                                                                    \
H
Hui Li 已提交
63
    char *_e = stpncpy(varDataVal(x), (str), (_maxs)-VARSTR_HEADER_SIZE); \
64
    varDataSetLen(x, (_e - (x)-VARSTR_HEADER_SIZE));                      \
H
Hongze Cheng 已提交
65 66
  } while (0)

67 68 69 70
#define STR_WITH_SIZE_TO_VARSTR(x, str, _size)  \
  do {                                          \
    *(VarDataLenT *)(x) = (VarDataLenT)(_size); \
    memcpy(varDataVal(x), (str), (_size));      \
H
Hongze Cheng 已提交
71
  } while (0);
H
hjxilinx 已提交
72

H
hzcheng 已提交
73 74
// ----------------- TSDB COLUMN DEFINITION
typedef struct {
C
Cary Xu 已提交
75 76 77
  int8_t   type;    // Column type
  int16_t  colId;   // column ID
  uint16_t bytes;   // column bytes
C
Cary Xu 已提交
78
  uint16_t offset;  // point offset in SDataRow after the header part.
H
hzcheng 已提交
79 80 81 82 83 84 85 86 87 88 89 90 91 92
} STColumn;

#define colType(col) ((col)->type)
#define colColId(col) ((col)->colId)
#define colBytes(col) ((col)->bytes)
#define colOffset(col) ((col)->offset)

#define colSetType(col, t) (colType(col) = (t))
#define colSetColId(col, id) (colColId(col) = (id))
#define colSetBytes(col, b) (colBytes(col) = (b))
#define colSetOffset(col, o) (colOffset(col) = (o))

// ----------------- TSDB SCHEMA DEFINITION
typedef struct {
H
Hongze Cheng 已提交
93
  int      version;    // version
H
hzcheng 已提交
94
  int      numOfCols;  // Number of columns appended
H
refact  
Hongze Cheng 已提交
95
  int      tlen;       // maximum length of a SDataRow without the header part (sizeof(VarDataOffsetT) + sizeof(VarDataLenT) + (bytes))
H
TD-353  
Hongze Cheng 已提交
96
  uint16_t flen;       // First part length in a SDataRow after the header part
H
refact  
Hongze Cheng 已提交
97
  uint16_t vlen;       // pure value part length, excluded the overhead (bytes only)
H
hzcheng 已提交
98 99 100 101
  STColumn columns[];
} STSchema;

#define schemaNCols(s) ((s)->numOfCols)
H
Hongze Cheng 已提交
102
#define schemaVersion(s) ((s)->version)
H
TD-166  
hzcheng 已提交
103 104
#define schemaTLen(s) ((s)->tlen)
#define schemaFLen(s) ((s)->flen)
T
Tao Liu 已提交
105
#define schemaVLen(s) ((s)->vlen)
H
hzcheng 已提交
106
#define schemaColAt(s, i) ((s)->columns + i)
S
TD-1848  
Shengliang Guan 已提交
107
#define tdFreeSchema(s) tfree((s))
H
hzcheng 已提交
108 109

STSchema *tdDupSchema(STSchema *pSchema);
H
TD-353  
Hongze Cheng 已提交
110
int       tdEncodeSchema(void **buf, STSchema *pSchema);
H
TD-353  
Hongze Cheng 已提交
111
void *    tdDecodeSchema(void *buf, STSchema **pRSchema);
H
hzcheng 已提交
112

H
Hongze Cheng 已提交
113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128
static FORCE_INLINE int comparColId(const void *key1, const void *key2) {
  if (*(int16_t *)key1 > ((STColumn *)key2)->colId) {
    return 1;
  } else if (*(int16_t *)key1 < ((STColumn *)key2)->colId) {
    return -1;
  } else {
    return 0;
  }
}

static FORCE_INLINE STColumn *tdGetColOfID(STSchema *pSchema, int16_t colId) {
  void *ptr = bsearch(&colId, (void *)pSchema->columns, schemaNCols(pSchema), sizeof(STColumn), comparColId);
  if (ptr == NULL) return NULL;
  return (STColumn *)ptr;
}

H
Hongze Cheng 已提交
129 130 131 132 133
// ----------------- SCHEMA BUILDER DEFINITION
typedef struct {
  int       tCols;
  int       nCols;
  int       tlen;
H
TD-353  
Hongze Cheng 已提交
134 135
  uint16_t  flen;
  uint16_t  vlen;
H
Hongze Cheng 已提交
136 137 138 139 140 141 142
  int       version;
  STColumn *columns;
} STSchemaBuilder;

int       tdInitTSchemaBuilder(STSchemaBuilder *pBuilder, int32_t version);
void      tdDestroyTSchemaBuilder(STSchemaBuilder *pBuilder);
void      tdResetTSchemaBuilder(STSchemaBuilder *pBuilder, int32_t version);
143
int       tdAddColToSchema(STSchemaBuilder *pBuilder, int8_t type, int16_t colId, int16_t bytes);
H
Hongze Cheng 已提交
144 145
STSchema *tdGetSchemaFromBuilder(STSchemaBuilder *pBuilder);

H
TD-1548  
Hongze Cheng 已提交
146 147 148 149 150
// ----------------- Semantic timestamp key definition
typedef uint64_t TKEY;

#define TKEY_INVALID UINT64_MAX
#define TKEY_NULL TKEY_INVALID
H
Hongze Cheng 已提交
151 152
#define TKEY_NEGATIVE_FLAG (((TKEY)1) << 63)
#define TKEY_DELETE_FLAG (((TKEY)1) << 62)
H
TD-1548  
Hongze Cheng 已提交
153 154 155 156 157 158 159 160
#define TKEY_VALUE_FILTER (~(TKEY_NEGATIVE_FLAG | TKEY_DELETE_FLAG))

#define TKEY_IS_NEGATIVE(tkey) (((tkey)&TKEY_NEGATIVE_FLAG) != 0)
#define TKEY_IS_DELETED(tkey) (((tkey)&TKEY_DELETE_FLAG) != 0)
#define tdSetTKEYDeleted(tkey) ((tkey) | TKEY_DELETE_FLAG)
#define tdGetTKEY(key) (((TKEY)ABS(key)) | (TKEY_NEGATIVE_FLAG & (TKEY)(key)))
#define tdGetKey(tkey) (((TSKEY)((tkey)&TKEY_VALUE_FILTER)) * (TKEY_IS_NEGATIVE(tkey) ? -1 : 1))

D
dapan1121 已提交
161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176
#define MIN_TS_KEY ((TSKEY)0x8000000000000001)
#define MAX_TS_KEY ((TSKEY)0x3fffffffffffffff)

#define TD_TO_TKEY(key) tdGetTKEY(((key) < MIN_TS_KEY) ? MIN_TS_KEY : (((key) > MAX_TS_KEY) ? MAX_TS_KEY : key))

static FORCE_INLINE TKEY keyToTkey(TSKEY key) {
  TSKEY lkey = key;
  if (key > MAX_TS_KEY) {
    lkey = MAX_TS_KEY;
  } else if (key < MIN_TS_KEY) {
    lkey = MIN_TS_KEY;
  }

  return tdGetTKEY(lkey);
}

H
TD-1548  
Hongze Cheng 已提交
177 178 179 180 181 182 183 184 185 186 187 188
static FORCE_INLINE int tkeyComparFn(const void *tkey1, const void *tkey2) {
  TSKEY key1 = tdGetKey(*(TKEY *)tkey1);
  TSKEY key2 = tdGetKey(*(TKEY *)tkey2);

  if (key1 < key2) {
    return -1;
  } else if (key1 > key2) {
    return 1;
  } else {
    return 0;
  }
}
H
hzcheng 已提交
189

C
Cary Xu 已提交
190 191 192
// ----------------- Data row structure

/* A data row, the format is like below:
K
kailixu 已提交
193
 * |<------------------------------------------------ len ---------------------------------->|
C
Cary Xu 已提交
194 195 196 197 198 199 200 201 202 203 204
 * |<--     Head      -->|<---------   flen -------------->|                                 |
 * +---------------------+---------------------------------+---------------------------------+
 * | uint16_t |  int16_t |                                 |                                 |
 * +----------+----------+---------------------------------+---------------------------------+
 * |   len    | sversion |           First part            |             Second part         |
 * +----------+----------+---------------------------------+---------------------------------+
 *
 * NOTE: timestamp in this row structure is TKEY instead of TSKEY
 */
typedef void *SDataRow;

B
Bomin Zhang 已提交
205
#define TD_DATA_ROW_HEAD_SIZE (sizeof(uint16_t) + sizeof(int16_t))
H
hzcheng 已提交
206

C
Cary Xu 已提交
207
#define dataRowLen(r) (*(TDRowLenT *)(r))  // 0~65535
C
Cary Xu 已提交
208
#define dataRowVersion(r) (*(int16_t *)POINTER_SHIFT(r, sizeof(int16_t)))
H
hzcheng 已提交
209
#define dataRowTuple(r) POINTER_SHIFT(r, TD_DATA_ROW_HEAD_SIZE)
H
TD-1548  
Hongze Cheng 已提交
210 211
#define dataRowTKey(r) (*(TKEY *)(dataRowTuple(r)))
#define dataRowKey(r) tdGetKey(dataRowTKey(r))
H
hzcheng 已提交
212
#define dataRowSetLen(r, l) (dataRowLen(r) = (l))
H
TD-90  
Hongze Cheng 已提交
213
#define dataRowSetVersion(r, v) (dataRowVersion(r) = (v))
H
hzcheng 已提交
214
#define dataRowCpy(dst, r) memcpy((dst), (r), dataRowLen(r))
H
TD-166  
hzcheng 已提交
215
#define dataRowMaxBytesFromSchema(s) (schemaTLen(s) + TD_DATA_ROW_HEAD_SIZE)
H
TD-1548  
Hongze Cheng 已提交
216
#define dataRowDeleted(r) TKEY_IS_DELETED(dataRowTKey(r))
H
hzcheng 已提交
217

C
Cary Xu 已提交
218 219
SDataRow tdNewDataRowFromSchema(STSchema *pSchema);
void     tdFreeDataRow(SDataRow row);
H
TD-166  
hzcheng 已提交
220
void     tdInitDataRow(SDataRow row, STSchema *pSchema);
C
Cary Xu 已提交
221
SDataRow tdDataRowDup(SDataRow row);
H
more  
Hongze Cheng 已提交
222

223
// offset here not include dataRow header length
C
Cary Xu 已提交
224
static FORCE_INLINE int tdAppendColVal(SDataRow row, const void *value, int8_t type, int32_t offset) {
225 226 227
  ASSERT(value != NULL);
  int32_t toffset = offset + TD_DATA_ROW_HEAD_SIZE;

H
TD-1548  
Hongze Cheng 已提交
228 229
  if (IS_VAR_DATA_TYPE(type)) {
    *(VarDataOffsetT *)POINTER_SHIFT(row, toffset) = dataRowLen(row);
K
kailixu 已提交
230
    memcpy(POINTER_SHIFT(row, dataRowLen(row)), value, varDataTLen(value));
H
TD-1548  
Hongze Cheng 已提交
231 232 233 234 235
    dataRowLen(row) += varDataTLen(value);
  } else {
    if (offset == 0) {
      ASSERT(type == TSDB_DATA_TYPE_TIMESTAMP);
      TKEY tvalue = tdGetTKEY(*(TSKEY *)value);
C
Cary Xu 已提交
236
      memcpy(POINTER_SHIFT(row, toffset), (const void *)(&tvalue), TYPE_BYTES[type]);
H
TD-1548  
Hongze Cheng 已提交
237
    } else {
238
      memcpy(POINTER_SHIFT(row, toffset), value, TYPE_BYTES[type]);
H
TD-1548  
Hongze Cheng 已提交
239
    }
240 241 242 243 244
  }

  return 0;
}

C
Cary Xu 已提交
245
// NOTE: offset here including the header size
C
Cary Xu 已提交
246
static FORCE_INLINE void *tdGetRowDataOfCol(SDataRow row, int8_t type, int32_t offset) {
C
Cary Xu 已提交
247 248 249 250 251 252 253
  if (IS_VAR_DATA_TYPE(type)) {
    return POINTER_SHIFT(row, *(VarDataOffsetT *)POINTER_SHIFT(row, offset));
  } else {
    return POINTER_SHIFT(row, offset);
  }
}

H
TD-34  
hzcheng 已提交
254 255
// ----------------- Data column structure
typedef struct SDataCol {
H
TD-166  
hzcheng 已提交
256 257 258
  int8_t          type;       // column type
  int16_t         colId;      // column ID
  int             bytes;      // column data bytes defined
H
TD-166  
hzcheng 已提交
259
  int             offset;     // data offset in a SDataRow (including the header size)
H
TD-166  
hzcheng 已提交
260 261 262 263
  int             spaceSize;  // Total space size for this column
  int             len;        // column data length
  VarDataOffsetT *dataOff;    // For binary and nchar data, the offset in the data column
  void *          pData;      // Actual data pointer
L
lichuang 已提交
264
  TSKEY           ts;         // only used in last NULL column
H
TD-34  
hzcheng 已提交
265 266
} SDataCol;

K
kailixu 已提交
267
#define isAllRowsNull(pCol) ((pCol)->len == 0)
H
TD-166  
hzcheng 已提交
268 269 270
static FORCE_INLINE void dataColReset(SDataCol *pDataCol) { pDataCol->len = 0; }

void dataColInit(SDataCol *pDataCol, STColumn *pCol, void **pBuf, int maxPoints);
K
kailixu 已提交
271
void dataColAppendVal(SDataCol *pCol, const void *value, int numOfRows, int maxPoints);
H
TD-166  
hzcheng 已提交
272 273
void dataColSetOffset(SDataCol *pCol, int nEle);

H
TD-166  
hzcheng 已提交
274 275
bool isNEleNull(SDataCol *pCol, int nEle);
void dataColSetNEleNull(SDataCol *pCol, int nEle, int maxPoints);
H
TD-166  
hzcheng 已提交
276

K
kailixu 已提交
277
const void *tdGetNullVal(int8_t type);
C
Cary Xu 已提交
278

H
TD-166  
hzcheng 已提交
279
// Get the data pointer from a column-wised data
K
kailixu 已提交
280 281 282 283
static FORCE_INLINE const void *tdGetColDataOfRow(SDataCol *pCol, int row) {
  if (isAllRowsNull(pCol)) {
    return tdGetNullVal(pCol->type);
  }
H
Hongze Cheng 已提交
284 285 286 287
  if (IS_VAR_DATA_TYPE(pCol->type)) {
    return POINTER_SHIFT(pCol->pData, pCol->dataOff[row]);
  } else {
    return POINTER_SHIFT(pCol->pData, TYPE_BYTES[pCol->type] * row);
H
TD-166  
hzcheng 已提交
288 289 290
  }
}

H
TD-166  
hzcheng 已提交
291
static FORCE_INLINE int32_t dataColGetNEleLen(SDataCol *pDataCol, int rows) {
H
TD-166  
hzcheng 已提交
292 293
  ASSERT(rows > 0);

H
Hongze Cheng 已提交
294 295 296 297
  if (IS_VAR_DATA_TYPE(pDataCol->type)) {
    return pDataCol->dataOff[rows - 1] + varDataTLen(tdGetColDataOfRow(pDataCol, rows - 1));
  } else {
    return TYPE_BYTES[pDataCol->type] * rows;
H
TD-166  
hzcheng 已提交
298 299 300
  }
}

H
TD-34  
hzcheng 已提交
301
typedef struct {
H
Hongze Cheng 已提交
302 303 304 305
  int maxRowSize;
  int maxCols;    // max number of columns
  int maxPoints;  // max number of points
  int bufSize;
H
TD-166  
hzcheng 已提交
306

H
Hongze Cheng 已提交
307 308 309 310 311
  int       numOfRows;
  int       numOfCols;  // Total number of cols
  int       sversion;   // TODO: set sversion
  void *    buf;
  SDataCol *cols;
H
TD-34  
hzcheng 已提交
312 313
} SDataCols;

H
TD-34  
hzcheng 已提交
314
#define keyCol(pCols) (&((pCols)->cols[0]))  // Key column
C
Cary Xu 已提交
315
#define dataColsTKeyAt(pCols, idx) ((TKEY *)(keyCol(pCols)->pData))[(idx)]  // the idx row of column-wised data
H
TD-1548  
Hongze Cheng 已提交
316
#define dataColsKeyAt(pCols, idx) tdGetKey(dataColsTKeyAt(pCols, idx))
317 318 319 320 321 322 323 324
static FORCE_INLINE TKEY dataColsTKeyFirst(SDataCols *pCols) {
  if (pCols->numOfRows) {
    return dataColsTKeyAt(pCols, 0);
  } else {
    return TKEY_INVALID;
  }
}

325 326 327 328 329
static FORCE_INLINE TSKEY dataColsKeyAtRow(SDataCols *pCols, int row) {
  ASSERT(row < pCols->numOfRows);
  return dataColsKeyAt(pCols, row);
}

330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352
static FORCE_INLINE TSKEY dataColsKeyFirst(SDataCols *pCols) {
  if (pCols->numOfRows) {
    return dataColsKeyAt(pCols, 0);
  } else {
    return TSDB_DATA_TIMESTAMP_NULL;
  }
}

static FORCE_INLINE TKEY dataColsTKeyLast(SDataCols *pCols) {
  if (pCols->numOfRows) {
    return dataColsTKeyAt(pCols, pCols->numOfRows - 1);
  } else {
    return TKEY_INVALID;
  }
}

static FORCE_INLINE TSKEY dataColsKeyLast(SDataCols *pCols) {
  if (pCols->numOfRows) {
    return dataColsKeyAt(pCols, pCols->numOfRows - 1);
  } else {
    return TSDB_DATA_TIMESTAMP_NULL;
  }
}
H
TD-34  
hzcheng 已提交
353

H
TD-166  
hzcheng 已提交
354
SDataCols *tdNewDataCols(int maxRowSize, int maxCols, int maxRows);
H
TD-34  
hzcheng 已提交
355
void       tdResetDataCols(SDataCols *pCols);
H
Hongze Cheng 已提交
356
int        tdInitDataCols(SDataCols *pCols, STSchema *pSchema);
H
TD-100  
hzcheng 已提交
357
SDataCols *tdDupDataCols(SDataCols *pCols, bool keepData);
H
Hongze Cheng 已提交
358
SDataCols *tdFreeDataCols(SDataCols *pCols);
H
Hongze Cheng 已提交
359
int        tdMergeDataCols(SDataCols *target, SDataCols *source, int rowsToMerge, int *pOffset);
H
more  
Hongze Cheng 已提交
360

H
Hongze Cheng 已提交
361
// ----------------- K-V data row structure
C
Cary Xu 已提交
362 363
/* |<-------------------------------------- len -------------------------------------------->|
 * |<----- header  ----->|<--------------------------- body -------------------------------->|
H
Hongze Cheng 已提交
364
 * +----------+----------+---------------------------------+---------------------------------+
C
Cary Xu 已提交
365
 * | uint16_t |  int16_t |                                 |                                 |
H
Hongze Cheng 已提交
366 367 368 369
 * +----------+----------+---------------------------------+---------------------------------+
 * |    len   |   ncols  |           cols index            |             data part           |
 * +----------+----------+---------------------------------+---------------------------------+
 */
H
Hongze Cheng 已提交
370
typedef void *SKVRow;
H
Hongze Cheng 已提交
371 372

typedef struct {
C
Cary Xu 已提交
373 374
  int16_t  colId;
  uint16_t offset;
H
Hongze Cheng 已提交
375 376
} SColIdx;

C
Cary Xu 已提交
377
#define TD_KV_ROW_HEAD_SIZE (sizeof(uint16_t) + sizeof(int16_t))
H
Hongze Cheng 已提交
378

C
Cary Xu 已提交
379
#define kvRowLen(r) (*(TDRowLenT *)(r))
C
Cary Xu 已提交
380
#define kvRowNCols(r) (*(int16_t *)POINTER_SHIFT(r, sizeof(uint16_t)))
H
Hongze Cheng 已提交
381 382
#define kvRowSetLen(r, len) kvRowLen(r) = (len)
#define kvRowSetNCols(r, n) kvRowNCols(r) = (n)
C
Cary Xu 已提交
383
#define kvRowColIdx(r) (SColIdx *)POINTER_SHIFT(r, TD_KV_ROW_HEAD_SIZE)
H
Hongze Cheng 已提交
384 385 386 387
#define kvRowValues(r) POINTER_SHIFT(r, TD_KV_ROW_HEAD_SIZE + sizeof(SColIdx) * kvRowNCols(r))
#define kvRowCpy(dst, r) memcpy((dst), (r), kvRowLen(r))
#define kvRowColVal(r, colIdx) POINTER_SHIFT(kvRowValues(r), (colIdx)->offset)
#define kvRowColIdxAt(r, i) (kvRowColIdx(r) + (i))
S
TD-1848  
Shengliang Guan 已提交
388
#define kvRowFree(r) tfree(r)
H
TD-90  
Hongze Cheng 已提交
389
#define kvRowEnd(r) POINTER_SHIFT(r, kvRowLen(r))
C
Cary Xu 已提交
390 391 392
#define kvRowTKey(r) (*(TKEY *)(kvRowValues(r)))
#define kvRowKey(r) tdGetKey(kvRowTKey(r))
#define kvRowDeleted(r) TKEY_IS_DELETED(kvRowTKey(r))
H
Hongze Cheng 已提交
393

H
Hongze Cheng 已提交
394
SKVRow tdKVRowDup(SKVRow row);
H
TD-90  
Hongze Cheng 已提交
395
int    tdSetKVRowDataOfCol(SKVRow *orow, int16_t colId, int8_t type, void *value);
H
TD-353  
Hongze Cheng 已提交
396
int    tdEncodeKVRow(void **buf, SKVRow row);
H
Hongze Cheng 已提交
397
void * tdDecodeKVRow(void *buf, SKVRow *row);
B
Bomin Zhang 已提交
398
void   tdSortKVRowByColIdx(SKVRow row);
H
Hongze Cheng 已提交
399 400 401 402 403 404 405 406 407 408 409

static FORCE_INLINE int comparTagId(const void *key1, const void *key2) {
  if (*(int16_t *)key1 > ((SColIdx *)key2)->colId) {
    return 1;
  } else if (*(int16_t *)key1 < ((SColIdx *)key2)->colId) {
    return -1;
  } else {
    return 0;
  }
}

H
Hongze Cheng 已提交
410
static FORCE_INLINE void *tdGetKVRowValOfCol(SKVRow row, int16_t colId) {
H
Hongze Cheng 已提交
411
  void *ret = taosbsearch(&colId, kvRowColIdx(row), kvRowNCols(row), sizeof(SColIdx), comparTagId, TD_EQ);
H
Hongze Cheng 已提交
412
  if (ret == NULL) return NULL;
H
Hongze Cheng 已提交
413
  return kvRowColVal(row, (SColIdx *)ret);
H
Hongze Cheng 已提交
414 415
}

L
liuyq-617 已提交
416 417 418 419
static FORCE_INLINE void *tdGetKVRowIdxOfCol(SKVRow row, int16_t colId) {
  return taosbsearch(&colId, kvRowColIdx(row), kvRowNCols(row), sizeof(SColIdx), comparTagId, TD_EQ);
}

C
Cary Xu 已提交
420 421 422 423 424 425 426 427
// offset here not include kvRow header length
static FORCE_INLINE int tdAppendKvColVal(SKVRow row, const void *value, int16_t colId, int8_t type, int32_t offset) {
  ASSERT(value != NULL);
  int32_t  toffset = offset + TD_KV_ROW_HEAD_SIZE;
  SColIdx *pColIdx = (SColIdx *)POINTER_SHIFT(row, toffset);
  char *   ptr = (char *)POINTER_SHIFT(row, kvRowLen(row));

  pColIdx->colId = colId;
C
Cary Xu 已提交
428
  pColIdx->offset = kvRowLen(row);  // offset of pColIdx including the TD_KV_ROW_HEAD_SIZE
C
Cary Xu 已提交
429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446

  if (IS_VAR_DATA_TYPE(type)) {
    memcpy(ptr, value, varDataTLen(value));
    kvRowLen(row) += varDataTLen(value);
  } else {
    if (offset == 0) {
      ASSERT(type == TSDB_DATA_TYPE_TIMESTAMP);
      TKEY tvalue = tdGetTKEY(*(TSKEY *)value);
      memcpy(ptr, (void *)(&tvalue), TYPE_BYTES[type]);
    } else {
      memcpy(ptr, value, TYPE_BYTES[type]);
    }
    kvRowLen(row) += TYPE_BYTES[type];
  }

  return 0;
}

H
Hongze Cheng 已提交
447 448 449 450 451
// ----------------- K-V data row builder
typedef struct {
  int16_t  tCols;
  int16_t  nCols;
  SColIdx *pColIdx;
C
Cary Xu 已提交
452 453
  uint16_t alloc;
  uint16_t size;
H
Hongze Cheng 已提交
454
  void *   buf;
H
Hongze Cheng 已提交
455
} SKVRowBuilder;
H
Hongze Cheng 已提交
456

H
Hongze Cheng 已提交
457 458 459 460
int    tdInitKVRowBuilder(SKVRowBuilder *pBuilder);
void   tdDestroyKVRowBuilder(SKVRowBuilder *pBuilder);
void   tdResetKVRowBuilder(SKVRowBuilder *pBuilder);
SKVRow tdGetKVRowFromBuilder(SKVRowBuilder *pBuilder);
H
Hongze Cheng 已提交
461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487

static FORCE_INLINE int tdAddColToKVRow(SKVRowBuilder *pBuilder, int16_t colId, int8_t type, void *value) {
  if (pBuilder->nCols >= pBuilder->tCols) {
    pBuilder->tCols *= 2;
    pBuilder->pColIdx = (SColIdx *)realloc((void *)(pBuilder->pColIdx), sizeof(SColIdx) * pBuilder->tCols);
    if (pBuilder->pColIdx == NULL) return -1;
  }

  pBuilder->pColIdx[pBuilder->nCols].colId = colId;
  pBuilder->pColIdx[pBuilder->nCols].offset = pBuilder->size;

  pBuilder->nCols++;

  int tlen = IS_VAR_DATA_TYPE(type) ? varDataTLen(value) : TYPE_BYTES[type];
  if (tlen > pBuilder->alloc - pBuilder->size) {
    while (tlen > pBuilder->alloc - pBuilder->size) {
      pBuilder->alloc *= 2;
    }
    pBuilder->buf = realloc(pBuilder->buf, pBuilder->alloc);
    if (pBuilder->buf == NULL) return -1;
  }

  memcpy(POINTER_SHIFT(pBuilder->buf, pBuilder->size), value, tlen);
  pBuilder->size += tlen;

  return 0;
}
H
Hongze Cheng 已提交
488

C
Cary Xu 已提交
489
// ----------------- SMemRow appended with sequential data row structure
C
Cary Xu 已提交
490
/*
C
Cary Xu 已提交
491
 * |---------|------------------------------------------------- len ---------------------------------->|
C
Cary Xu 已提交
492 493 494 495 496 497
 * |<--------     Head      ------>|<---------   flen -------------->|                                 |
 * |---------+---------------------+---------------------------------+---------------------------------+
 * | uint8_t | uint16_t |  int16_t |                                 |                                 |
 * |---------+----------+----------+---------------------------------+---------------------------------+
 * |  flag   |   len    | sversion |           First part            |             Second part         |
 * +---------+----------+----------+---------------------------------+---------------------------------+
C
Cary Xu 已提交
498 499 500
 *
 * NOTE: timestamp in this row structure is TKEY instead of TSKEY
 */
C
Cary Xu 已提交
501

C
Cary Xu 已提交
502
// ----------------- SMemRow appended with extended K-V data row structure
C
Cary Xu 已提交
503 504
/* |--------------------|------------------------------------------------  len ---------------------------------->|
 * |<-------------     Head      ------------>|<---------   flen -------------->|                                 |
C
Cary Xu 已提交
505 506 507 508 509
 * |--------------------+----------+--------------------------------------------+---------------------------------+
 * | uint8_t | int16_t  | uint16_t |  int16_t |                                 |                                 |
 * |---------+----------+----------+----------+---------------------------------+---------------------------------+
 * |   flag  | sversion |   len    |   ncols  |           cols index            |             data part           |
 * |---------+----------+----------+----------+---------------------------------+---------------------------------+
C
Cary Xu 已提交
510 511
 */

C
Cary Xu 已提交
512 513
typedef void *SMemRow;

C
Cary Xu 已提交
514
#define TD_MEM_ROW_TYPE_SIZE sizeof(uint8_t)
C
Cary Xu 已提交
515 516 517
#define TD_MEM_ROW_KV_VER_SIZE sizeof(int16_t)
#define TD_MEM_ROW_KV_TYPE_VER_SIZE (TD_MEM_ROW_TYPE_SIZE + TD_MEM_ROW_KV_VER_SIZE)
#define TD_MEM_ROW_DATA_HEAD_SIZE (TD_MEM_ROW_TYPE_SIZE + TD_DATA_ROW_HEAD_SIZE)
C
Cary Xu 已提交
518
// #define TD_MEM_ROW_KV_HEAD_SIZE (TD_MEM_ROW_TYPE_SIZE + TD_MEM_ROW_KV_VER_SIZE + TD_KV_ROW_HEAD_SIZE)
C
Cary Xu 已提交
519 520 521 522 523 524 525 526

#define SMEM_ROW_DATA 0U  // SDataRow
#define SMEM_ROW_KV 1U    // SKVRow

#define memRowType(r) (*(uint8_t *)(r))
#define isDataRow(r) (SMEM_ROW_DATA == memRowType(r))
#define isKvRow(r) (SMEM_ROW_KV == memRowType(r))

C
Cary Xu 已提交
527 528
#define memRowDataBody(r) POINTER_SHIFT(r, TD_MEM_ROW_TYPE_SIZE)  // section after flag
#define memRowKvBody(r) \
C
Cary Xu 已提交
529 530 531 532
  POINTER_SHIFT(r, TD_MEM_ROW_KV_TYPE_VER_SIZE)  // section after flag + sversion as to reuse SKVRow

#define memRowDataLen(r) (*(TDRowLenT *)memRowDataBody(r))  //  0~65535
#define memRowKvLen(r) (*(TDRowLenT *)memRowKvBody(r))      //  0~65535
C
Cary Xu 已提交
533

K
kailixu 已提交
534 535
#define memRowDataTLen(r) \
  ((TDRowTLenT)(memRowDataLen(r) + TD_MEM_ROW_TYPE_SIZE))  // using uint32_t/int32_t to store the TLen
C
Cary Xu 已提交
536

K
kailixu 已提交
537
#define memRowKvTLen(r) ((TDRowTLenT)(memRowKvLen(r) + TD_MEM_ROW_KV_TYPE_VER_SIZE))
C
Cary Xu 已提交
538 539

#define memRowLen(r) (isDataRow(r) ? memRowDataLen(r) : memRowKvLen(r))
C
Cary Xu 已提交
540
#define memRowTLen(r) (isDataRow(r) ? memRowDataTLen(r) : memRowKvTLen(r))  // using uint32_t/int32_t to store the TLen
C
Cary Xu 已提交
541 542 543 544

#define memRowDataVersion(r) dataRowVersion(memRowDataBody(r))
#define memRowKvVersion(r) (*(int16_t *)POINTER_SHIFT(r, TD_MEM_ROW_TYPE_SIZE))
#define memRowVersion(r) (isDataRow(r) ? memRowDataVersion(r) : memRowKvVersion(r))  // schema version
C
Cary Xu 已提交
545
#define memRowSetKvVersion(r, v) (memRowKvVersion(r) = (v))
C
Cary Xu 已提交
546
#define memRowTuple(r) (isDataRow(r) ? dataRowTuple(memRowDataBody(r)) : kvRowValues(memRowKvBody(r)))
C
Cary Xu 已提交
547

C
Cary Xu 已提交
548 549
#define memRowTKey(r) (isDataRow(r) ? dataRowTKey(memRowDataBody(r)) : kvRowTKey(memRowKvBody(r)))
#define memRowKey(r) (isDataRow(r) ? dataRowKey(memRowDataBody(r)) : kvRowKey(memRowKvBody(r)))
C
Cary Xu 已提交
550 551

#define memRowSetType(r, t) (memRowType(r) = (t))
C
Cary Xu 已提交
552 553
#define memRowSetLen(r, l) (isDataRow(r) ? memRowDataLen(r) = (l) : memRowKvLen(r) = (l))
#define memRowSetVersion(r, v) (isDataRow(r) ? dataRowSetVersion(memRowDataBody(r), v) : memRowKvSetVersion(r, v))
C
Cary Xu 已提交
554
#define memRowCpy(dst, r) memcpy((dst), (r), memRowTLen(r))
C
Cary Xu 已提交
555
#define memRowMaxBytesFromSchema(s) (schemaTLen(s) + TD_MEM_ROW_DATA_HEAD_SIZE)
C
Cary Xu 已提交
556 557
#define memRowDeleted(r) TKEY_IS_DELETED(memRowTKey(r))

C
Cary Xu 已提交
558 559
SMemRow tdMemRowDup(SMemRow row);
void    tdAppendMemRowToDataCol(SMemRow row, STSchema *pSchema, SDataCols *pCols);
C
Cary Xu 已提交
560
// NOTE: offset here including the header size
L
liuyq-617 已提交
561
static FORCE_INLINE void *tdGetKvRowDataOfCol(void *row, int32_t offset) { return POINTER_SHIFT(row, offset); }
C
Cary Xu 已提交
562
// NOTE: offset here including the header size
C
Cary Xu 已提交
563 564 565 566
static FORCE_INLINE void *tdGetMemRowDataOfCol(void *row, int8_t type, int32_t offset) {
  if (isDataRow(row)) {
    return tdGetRowDataOfCol(row, type, offset);
  } else if (isKvRow(row)) {
L
liuyq-617 已提交
567
    return tdGetKvRowDataOfCol(row, offset);
C
Cary Xu 已提交
568 569 570 571 572 573
  } else {
    ASSERT(0);
  }
  return NULL;
}

C
Cary Xu 已提交
574 575 576 577 578 579 580 581 582 583
// ----------------- Raw payload structure for row:
/* |<------------ Head ------------->|<----------- body of column data tuple ------------------->|
 * |                                 |<----------------- flen ------------->|<--- value part --->|
 * |SMemRowType| dataTLen |  nCols   |  colId  | colType | offset   |  ...  | value |...|...|... |
 * +-----------+----------+----------+--------------------------------------|--------------------|
 * | uint8_t   | uint32_t | uint16_t | int16_t | uint8_t | uint16_t |  ...  |.......|...|...|... |
 * +-----------+----------+----------+--------------------------------------+--------------------|
 *  1. offset in column data tuple starts from the value part in case of uint16_t overflow.
 *  2. dataTLen: total length including the header and body.
 */
C
Cary Xu 已提交
584 585

#define PAYLOAD_NCOLS_LEN sizeof(uint16_t)
C
Cary Xu 已提交
586
#define PAYLOAD_NCOLS_OFFSET (sizeof(uint8_t) + sizeof(TDRowTLenT))
C
Cary Xu 已提交
587 588 589
#define PAYLOAD_HEADER_LEN (PAYLOAD_NCOLS_OFFSET + PAYLOAD_NCOLS_LEN)
#define PAYLOAD_ID_LEN sizeof(int16_t)
#define PAYLOAD_ID_TYPE_LEN (sizeof(int16_t) + sizeof(uint8_t))
C
Cary Xu 已提交
590
#define PAYLOAD_COL_HEAD_LEN (PAYLOAD_ID_TYPE_LEN + sizeof(uint16_t))
C
Cary Xu 已提交
591
#define PAYLOAD_PRIMARY_COL_LEN (PAYLOAD_ID_TYPE_LEN + sizeof(TSKEY))
C
Cary Xu 已提交
592 593 594 595

#define payloadBody(r) POINTER_SHIFT(r, PAYLOAD_HEADER_LEN)
#define payloadType(r) (*(uint8_t *)(r))
#define payloadSetType(r, t) (payloadType(r) = (t))
C
Cary Xu 已提交
596
#define payloadTLen(r) (*(TDRowTLenT *)POINTER_SHIFT(r, TD_MEM_ROW_TYPE_SIZE))  // including total header
C
Cary Xu 已提交
597 598 599
#define payloadSetTLen(r, l) (payloadTLen(r) = (l))
#define payloadNCols(r) (*(TDRowLenT *)POINTER_SHIFT(r, PAYLOAD_NCOLS_OFFSET))
#define payloadSetNCols(r, n) (payloadNCols(r) = (n))
C
Cary Xu 已提交
600 601 602 603 604 605 606 607 608 609 610 611 612 613
#define payloadValuesOffset(r) \
  (PAYLOAD_HEADER_LEN + payloadNCols(r) * PAYLOAD_COL_HEAD_LEN)    // avoid using the macro in loop
#define payloadValues(r) POINTER_SHIFT(r, payloadValuesOffset(r))  // avoid using the macro in loop
#define payloadColId(c) (*(int16_t *)(c))
#define payloadColType(c) (*(uint8_t *)POINTER_SHIFT(c, PAYLOAD_ID_LEN))
#define payloadColOffset(c) (*(uint16_t *)POINTER_SHIFT(c, PAYLOAD_ID_TYPE_LEN))
#define payloadColValue(c) POINTER_SHIFT(c, payloadColOffset(c))

#define payloadColSetId(c, i) (payloadColId(c) = (i))
#define payloadColSetType(c, t) (payloadColType(c) = (t))
#define payloadColSetOffset(c, o) (payloadColOffset(c) = (o))

#define payloadKeyOffset(r) (*(uint16_t *)POINTER_SHIFT(r, PAYLOAD_HEADER_LEN + PAYLOAD_ID_TYPE_LEN))
#define payloadTKey(r) (*(TKEY *)POINTER_SHIFT(r, payloadValuesOffset(r) + payloadKeyOffset(r)))
C
Cary Xu 已提交
614 615
#define payloadKey(r) tdGetKey(payloadTKey(r))

C
Cary Xu 已提交
616
static FORCE_INLINE char *payloadNextCol(char *pCol) { return (char *)POINTER_SHIFT(pCol, PAYLOAD_COL_HEAD_LEN); }
C
Cary Xu 已提交
617

H
more  
hzcheng 已提交
618 619 620 621
#ifdef __cplusplus
}
#endif

C
Cary Xu 已提交
622
#endif  // _TD_DATA_FORMAT_H_