tdataformat.h 7.0 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 18
#define _TD_DATA_FORMAT_H_

#include <stdint.h>
H
hzcheng 已提交
19
#include <stdlib.h>
H
hzcheng 已提交
20
#include <string.h>
H
more  
Hongze Cheng 已提交
21

H
hzcheng 已提交
22
#include "taosdef.h"
H
TD-166  
hzcheng 已提交
23
#include "tutil.h"
H
hzcheng 已提交
24

H
more  
hzcheng 已提交
25 26 27
#ifdef __cplusplus
extern "C" {
#endif
H
hzcheng 已提交
28 29 30 31 32 33

// ----------------- TSDB COLUMN DEFINITION
typedef struct {
  int8_t  type;    // Column type
  int16_t colId;   // column ID
  int32_t bytes;   // column bytes
H
TD-166  
hzcheng 已提交
34
  int32_t offset;  // point offset in SDataRow after the header part
H
hzcheng 已提交
35 36 37 38 39 40 41 42 43 44 45 46 47 48
} 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
TD-166  
hzcheng 已提交
49
  int      totalCols;  // Total columns allocated
H
hzcheng 已提交
50
  int      numOfCols;  // Number of columns appended
H
TD-166  
hzcheng 已提交
51 52
  int      tlen;       // maximum length of a SDataRow without the header part
  int      flen;       // First part length in a SDataRow after the header part
H
hzcheng 已提交
53 54 55 56
  STColumn columns[];
} STSchema;

#define schemaNCols(s) ((s)->numOfCols)
H
TD-166  
hzcheng 已提交
57 58 59
#define schemaTotalCols(s) ((s)->totalCols)
#define schemaTLen(s) ((s)->tlen)
#define schemaFLen(s) ((s)->flen)
H
hzcheng 已提交
60 61 62
#define schemaColAt(s, i) ((s)->columns + i)

STSchema *tdNewSchema(int32_t nCols);
H
TD-166  
hzcheng 已提交
63 64
#define   tdFreeSchema(s) tfree((s))
int       tdSchemaAddCol(STSchema *pSchema, int8_t type, int16_t colId, int32_t bytes);
H
hzcheng 已提交
65
STSchema *tdDupSchema(STSchema *pSchema);
H
TD-27  
hzcheng 已提交
66 67 68
int       tdGetSchemaEncodeSize(STSchema *pSchema);
void *    tdEncodeSchema(void *dst, STSchema *pSchema);
STSchema *tdDecodeSchema(void **psrc);
H
hzcheng 已提交
69

H
more  
Hongze Cheng 已提交
70 71
// ----------------- Data row structure

H
hzcheng 已提交
72
/* A data row, the format is like below:
H
TD-166  
hzcheng 已提交
73 74 75 76 77 78 79
 * |<------------------------------------- len ---------------------------------->|
 * |<--Head ->|<---------   flen -------------->|                                 |
 * +----------+---------------------------------+---------------------------------+
 * | int32_t  |                                 |                                 |
 * +----------+---------------------------------+---------------------------------+
 * |   len    |           First part            |             Second part         |
 * +----------+---------------------------------+---------------------------------+
H
more  
Hongze Cheng 已提交
80
 */
H
hzcheng 已提交
81 82
typedef void *SDataRow;

H
TD-166  
hzcheng 已提交
83
#define TD_DATA_ROW_HEAD_SIZE sizeof(int32_t)
H
hzcheng 已提交
84

H
hzcheng 已提交
85
#define dataRowLen(r) (*(int32_t *)(r))
H
TD-166  
hzcheng 已提交
86
#define dataRowTuple(r) POINTER_DRIFT(r, TD_DATA_ROW_HEAD_SIZE)
H
TD-34  
hzcheng 已提交
87
#define dataRowKey(r) (*(TSKEY *)(dataRowTuple(r)))
H
hzcheng 已提交
88
#define dataRowSetLen(r, l) (dataRowLen(r) = (l))
H
hzcheng 已提交
89
#define dataRowCpy(dst, r) memcpy((dst), (r), dataRowLen(r))
H
TD-166  
hzcheng 已提交
90
#define dataRowMaxBytesFromSchema(s) (schemaTLen(s) + TD_DATA_ROW_HEAD_SIZE)
H
hzcheng 已提交
91

H
hzcheng 已提交
92
SDataRow tdNewDataRowFromSchema(STSchema *pSchema);
H
hzcheng 已提交
93
void     tdFreeDataRow(SDataRow row);
H
TD-166  
hzcheng 已提交
94
void     tdInitDataRow(SDataRow row, STSchema *pSchema);
H
TD-166  
hzcheng 已提交
95
int      tdAppendColVal(SDataRow row, void *value, int8_t type, int32_t bytes, int32_t offset);
H
hzcheng 已提交
96
SDataRow tdDataRowDup(SDataRow row);
H
more  
Hongze Cheng 已提交
97

H
TD-166  
hzcheng 已提交
98
// NOTE: offset here including the header size
H
TD-166  
hzcheng 已提交
99 100 101 102
static FORCE_INLINE void *tdGetRowDataOfCol(SDataRow row, int8_t type, int32_t offset) {
  switch (type) {
    case TSDB_DATA_TYPE_BINARY:
    case TSDB_DATA_TYPE_NCHAR:
H
TD-166  
hzcheng 已提交
103
      return POINTER_DRIFT(row, *(VarDataOffsetT *)POINTER_DRIFT(row, offset));
H
TD-166  
hzcheng 已提交
104 105
      break;
    default:
H
TD-166  
hzcheng 已提交
106
      return POINTER_DRIFT(row, offset);
H
TD-166  
hzcheng 已提交
107 108 109 110
      break;
  }
}

H
TD-34  
hzcheng 已提交
111 112
// ----------------- Data column structure
typedef struct SDataCol {
H
TD-166  
hzcheng 已提交
113 114 115
  int8_t          type;       // column type
  int16_t         colId;      // column ID
  int             bytes;      // column data bytes defined
H
TD-166  
hzcheng 已提交
116
  int             offset;     // data offset in a SDataRow (including the header size)
H
TD-166  
hzcheng 已提交
117 118 119 120
  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
H
TD-34  
hzcheng 已提交
121 122
} SDataCol;

H
TD-166  
hzcheng 已提交
123 124 125
static FORCE_INLINE void dataColReset(SDataCol *pDataCol) { pDataCol->len = 0; }

void dataColInit(SDataCol *pDataCol, STColumn *pCol, void **pBuf, int maxPoints);
H
TD-166  
hzcheng 已提交
126
void dataColAppendVal(SDataCol *pCol, void *value, int numOfPoints, int maxPoints);
H
TD-166  
hzcheng 已提交
127 128 129
void dataColPopPoints(SDataCol *pCol, int pointsToPop, int numOfPoints);
void dataColSetOffset(SDataCol *pCol, int nEle);

H
TD-166  
hzcheng 已提交
130 131
bool isNEleNull(SDataCol *pCol, int nEle);
void dataColSetNEleNull(SDataCol *pCol, int nEle, int maxPoints);
H
TD-166  
hzcheng 已提交
132 133 134

// Get the data pointer from a column-wised data
static FORCE_INLINE void *tdGetColDataOfRow(SDataCol *pCol, int row) {
H
TD-166  
hzcheng 已提交
135 136 137 138 139 140 141 142 143
  switch (pCol->type) {
    case TSDB_DATA_TYPE_BINARY:
    case TSDB_DATA_TYPE_NCHAR:
      return POINTER_DRIFT(pCol->pData, pCol->dataOff[row]);
      break;

    default:
      return POINTER_DRIFT(pCol->pData, TYPE_BYTES[pCol->type] * row);
      break;
H
TD-166  
hzcheng 已提交
144 145 146
  }
}

H
TD-166  
hzcheng 已提交
147
static FORCE_INLINE int32_t dataColGetNEleLen(SDataCol *pDataCol, int rows) {
H
TD-166  
hzcheng 已提交
148 149
  ASSERT(rows > 0);

H
TD-166  
hzcheng 已提交
150 151 152
  switch (pDataCol->type) {
    case TSDB_DATA_TYPE_BINARY:
    case TSDB_DATA_TYPE_NCHAR:
H
TD-166  
hzcheng 已提交
153
      return pDataCol->dataOff[rows - 1] + varDataTLen(tdGetColDataOfRow(pDataCol, rows - 1));
H
TD-166  
hzcheng 已提交
154 155
      break;
    default:
H
TD-166  
hzcheng 已提交
156
      return TYPE_BYTES[pDataCol->type] * rows;
H
TD-166  
hzcheng 已提交
157 158 159 160
  }
}


H
TD-34  
hzcheng 已提交
161
typedef struct {
H
TD-34  
hzcheng 已提交
162 163 164
  int      maxRowSize;
  int      maxCols;    // max number of columns
  int      maxPoints;  // max number of points
H
TD-166  
hzcheng 已提交
165
  int      bufSize;
H
TD-166  
hzcheng 已提交
166

H
TD-34  
hzcheng 已提交
167
  int      numOfPoints;
H
TD-34  
hzcheng 已提交
168
  int      numOfCols;  // Total number of cols
H
TD-34  
hzcheng 已提交
169
  int      sversion;   // TODO: set sversion
H
TD-34  
hzcheng 已提交
170 171 172 173
  void *   buf;
  SDataCol cols[];
} SDataCols;

H
TD-34  
hzcheng 已提交
174
#define keyCol(pCols) (&((pCols)->cols[0]))  // Key column
H
TD-166  
hzcheng 已提交
175
#define dataColsKeyAt(pCols, idx) ((TSKEY *)(keyCol(pCols)->pData))[(idx)]
H
TD-34  
hzcheng 已提交
176 177
#define dataColsKeyFirst(pCols) dataColsKeyAt(pCols, 0)
#define dataColsKeyLast(pCols) dataColsKeyAt(pCols, (pCols)->numOfPoints - 1)
H
TD-34  
hzcheng 已提交
178

H
TD-166  
hzcheng 已提交
179
SDataCols *tdNewDataCols(int maxRowSize, int maxCols, int maxRows);
H
TD-34  
hzcheng 已提交
180
void       tdResetDataCols(SDataCols *pCols);
H
TD-34  
hzcheng 已提交
181
void       tdInitDataCols(SDataCols *pCols, STSchema *pSchema);
H
TD-100  
hzcheng 已提交
182
SDataCols *tdDupDataCols(SDataCols *pCols, bool keepData);
H
TD-34  
hzcheng 已提交
183
void       tdFreeDataCols(SDataCols *pCols);
H
TD-34  
hzcheng 已提交
184
void       tdAppendDataRowToDataCol(SDataRow row, SDataCols *pCols);
H
TD-34  
hzcheng 已提交
185
void       tdPopDataColsPoints(SDataCols *pCols, int pointsToPop);
H
hzcheng 已提交
186
int        tdMergeDataCols(SDataCols *target, SDataCols *src, int rowsToMerge);
H
TD-100  
hzcheng 已提交
187
void       tdMergeTwoDataCols(SDataCols *target, SDataCols *src1, int *iter1, SDataCols *src2, int *iter2, int tRows);
H
more  
Hongze Cheng 已提交
188

H
more  
hzcheng 已提交
189 190 191 192
#ifdef __cplusplus
}
#endif

H
hzcheng 已提交
193
#endif  // _TD_DATA_FORMAT_H_