/
TimestampColumnWriter.go
320 lines (289 loc) · 12.4 KB
/
TimestampColumnWriter.go
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
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
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
package store
import (
"fmt"
"time"
"github.com/mothdb-bd/orc-go/pkg/function"
"github.com/mothdb-bd/orc-go/pkg/maths"
"github.com/mothdb-bd/orc-go/pkg/optional"
"github.com/mothdb-bd/orc-go/pkg/spi/block"
"github.com/mothdb-bd/orc-go/pkg/store/metadata"
"github.com/mothdb-bd/orc-go/pkg/util"
)
var (
TIMESTAMP_INSTANCE_SIZE int32 = util.SizeOf(&TimestampColumnWriter{})
TIMESTAMP_MOTH_EPOCH_IN_SECONDS int64 = int64(time.Date(2015, 1, 1, 0, 0, 0, 0, time.UTC).Second())
)
// type TimestampKind int8
// const (
// TIMESTAMP_MILLIS TimestampKind = iota
// TIMESTAMP_MICROS
// TIMESTAMP_NANOS
// INSTANT_MILLIS
// INSTANT_MICROS
// INSTANT_NANOS
// )
// The MOTH encoding erroneously uses normal integer division to compute seconds,
// rather than floor modulus, which produces the wrong result for negative values
// (those that are before the epoch). Readers must correct for this. It also makes
// it impossible to represent values less than one second before the epoch, which
// must also be handled in MothWriteValidation.
//
// The sub-second value (nanoseconds) typically has a large number of trailing zeroes,
// as many systems only record millisecond or microsecond precision. To optimize storage,
// if the value has at least two trailing zeros, the trailing decimal zero digits are
// removed, and the last three bits record how many zeros were removed, minus one:
//
// # Trailing 0s Last 3 Bits Example nanos Example encoding
// 0 0b000 123456789 (123456789 << 3) | 0b000
// 1 0b000 123456780 (123456780 << 3) | 0b000
// 2 0b001 123456700 (1234567 << 3) | 0b001
// 3 0b010 123456000 (123456 << 3) | 0b010
// 4 0b011 123450000 (12345 << 3) | 0b011
// 5 0b100 123400000 (1234 << 3) | 0b100
// 6 0b101 123000000 (123 << 3) | 0b101
// 7 0b110 120000000 (12 << 3) | 0b110
// 8 0b111 100000000 (1 << 3) | 0b111
type TimestampColumnWriter struct {
//继承
ColumnWriter
columnId metadata.MothColumnId
kind block.Type
timestampKind TimestampKind
compressed bool
columnEncoding *metadata.ColumnEncoding
secondsStream LongOutputStream
nanosStream LongOutputStream
presentStream *PresentOutputStream
rowGroupColumnStatistics *util.ArrayList[*metadata.ColumnStatistics]
statisticsBuilderSupplier function.Supplier[*metadata.TimestampStatisticsBuilder]
statisticsBuilder metadata.LongValueStatisticsBuilder
closed bool
}
func NewTimestampColumnWriter(columnId metadata.MothColumnId, kind block.Type, compression metadata.CompressionKind, bufferSize int32, statisticsBuilderSupplier function.Supplier[*metadata.TimestampStatisticsBuilder]) *TimestampColumnWriter {
tr := new(TimestampColumnWriter)
tr.columnId = columnId
tr.kind = kind
tr.timestampKind = timestampKindForType(kind)
tr.compressed = compression != metadata.NONE
tr.columnEncoding = metadata.NewColumnEncoding(metadata.DIRECT_V2, 0)
tr.secondsStream = NewLongOutputStreamV2(compression, bufferSize, true, metadata.DATA)
tr.nanosStream = NewLongOutputStreamV2(compression, bufferSize, false, metadata.SECONDARY)
tr.presentStream = NewPresentOutputStream(compression, bufferSize)
tr.statisticsBuilderSupplier = statisticsBuilderSupplier
tr.statisticsBuilder = statisticsBuilderSupplier.Get()
tr.rowGroupColumnStatistics = util.NewArrayList[*metadata.ColumnStatistics]()
return tr
}
func timestampKindForType(kind block.Type) TimestampKind {
if kind.Equals(block.TIMESTAMP_MILLIS) {
return TIMESTAMP_MILLIS
}
if kind.Equals(block.TIMESTAMP_MICROS) {
return TIMESTAMP_MICROS
}
if kind.Equals(block.TIMESTAMP_NANOS) {
return TIMESTAMP_NANOS
}
if kind.Equals(block.TIMESTAMP_TZ_MILLIS) {
return INSTANT_MILLIS
}
if kind.Equals(block.TIMESTAMP_TZ_MICROS) {
return INSTANT_MICROS
}
if kind.Equals(block.TIMESTAMP_TZ_NANOS) {
return INSTANT_NANOS
}
panic(fmt.Sprintf("Unsupported kind for MOTH timestamp writer: %s", kind))
}
// @Override
func (tr *TimestampColumnWriter) GetNestedColumnWriters() *util.ArrayList[ColumnWriter] {
return util.NewArrayList[ColumnWriter]()
}
// @Override
func (tr *TimestampColumnWriter) GetColumnEncodings() map[metadata.MothColumnId]*metadata.ColumnEncoding {
return util.NewMap(tr.columnId, tr.columnEncoding)
}
// @Override
func (tr *TimestampColumnWriter) BeginRowGroup() {
tr.presentStream.RecordCheckpoint()
tr.secondsStream.RecordCheckpoint()
tr.nanosStream.RecordCheckpoint()
}
// @Override
func (tr *TimestampColumnWriter) WriteBlock(block block.Block) {
util.CheckState(!tr.closed)
util.CheckArgument2(block.GetPositionCount() > 0, "Block is empty")
for position := util.INT32_ZERO; position < block.GetPositionCount(); position++ {
tr.presentStream.WriteBoolean(!block.IsNull(position))
}
switch tr.timestampKind {
case TIMESTAMP_MILLIS, TIMESTAMP_MICROS:
tr.writeTimestampMicros(block)
case TIMESTAMP_NANOS:
tr.writeTimestampNanos(block)
case INSTANT_MILLIS:
tr.writeInstantShort(block)
case INSTANT_MICROS, INSTANT_NANOS:
tr.writeInstantLong(block)
}
}
// @Override
func (tr *TimestampColumnWriter) FinishRowGroup() map[metadata.MothColumnId]*metadata.ColumnStatistics {
util.CheckState(!tr.closed)
statistics := tr.statisticsBuilder.BuildColumnStatistics()
tr.rowGroupColumnStatistics.Add(statistics)
tr.statisticsBuilder = tr.statisticsBuilderSupplier.Get()
return util.NewMap(tr.columnId, statistics)
}
// @Override
func (tr *TimestampColumnWriter) Close() {
tr.closed = true
tr.secondsStream.Close()
tr.nanosStream.Close()
tr.presentStream.Close()
}
// @Override
func (tr *TimestampColumnWriter) GetColumnStripeStatistics() map[metadata.MothColumnId]*metadata.ColumnStatistics {
util.CheckState(tr.closed)
return util.NewMap(tr.columnId, metadata.MergeColumnStatistics(tr.rowGroupColumnStatistics))
}
// @Override
func (tr *TimestampColumnWriter) GetIndexStreams(metadataWriter *CompressedMetadataWriter) *util.ArrayList[*StreamDataOutput] {
util.CheckState(tr.closed)
rowGroupIndexes := util.NewArrayList[*metadata.RowGroupIndex]()
secondsCheckpoints := tr.secondsStream.GetCheckpoints()
nanosCheckpoints := tr.nanosStream.GetCheckpoints()
presentCheckpoints := tr.presentStream.GetCheckpoints()
for i := util.INT32_ZERO; i < tr.rowGroupColumnStatistics.SizeInt32(); i++ {
groupId := i
columnStatistics := tr.rowGroupColumnStatistics.GetByInt32(groupId)
secondsCheckpoint := secondsCheckpoints.GetByInt32(groupId)
nanosCheckpoint := nanosCheckpoints.GetByInt32(groupId)
presentCheckpoint := optional.Map(presentCheckpoints, func(checkpoints *util.ArrayList[*BooleanStreamCheckpoint]) *BooleanStreamCheckpoint {
return checkpoints.GetByInt32(groupId)
})
positions := createTimestampColumnPositionList(tr.compressed, secondsCheckpoint, nanosCheckpoint, presentCheckpoint)
rowGroupIndexes.Add(metadata.NewRowGroupIndex(positions, columnStatistics))
}
slice := metadataWriter.WriteRowIndexes(rowGroupIndexes)
stream := metadata.NewStream(tr.columnId, metadata.ROW_INDEX, slice.SizeInt32(), false)
return util.NewArrayList(NewStreamDataOutput(slice, stream))
}
func createTimestampColumnPositionList(compressed bool, secondsCheckpoint LongStreamCheckpoint, nanosCheckpoint LongStreamCheckpoint, presentCheckpoint *optional.Optional[*BooleanStreamCheckpoint]) *util.ArrayList[int32] {
positionList := util.NewArrayList[int32]()
presentCheckpoint.IfPresent(func(booleanStreamCheckpoint *BooleanStreamCheckpoint) {
positionList.AddAll(booleanStreamCheckpoint.ToPositionList(compressed))
})
positionList.AddAll(secondsCheckpoint.ToPositionList(compressed))
positionList.AddAll(nanosCheckpoint.ToPositionList(compressed))
return positionList
}
// @Override
func (tr *TimestampColumnWriter) GetBloomFilters(metadataWriter *CompressedMetadataWriter) *util.ArrayList[*StreamDataOutput] {
return util.EMPTY_LIST[*StreamDataOutput]()
}
// @Override
func (tr *TimestampColumnWriter) GetDataStreams() *util.ArrayList[*StreamDataOutput] {
util.CheckState(tr.closed)
outputDataStreams := util.NewArrayList[*StreamDataOutput]()
tr.presentStream.GetStreamDataOutput(tr.columnId).IfPresent(
func(s *StreamDataOutput) {
outputDataStreams.Add(s)
})
outputDataStreams.Add(tr.secondsStream.GetStreamDataOutput(tr.columnId))
outputDataStreams.Add(tr.nanosStream.GetStreamDataOutput(tr.columnId))
return outputDataStreams
}
// @Override
func (tr *TimestampColumnWriter) GetBufferedBytes() int64 {
return tr.secondsStream.GetBufferedBytes() + tr.nanosStream.GetBufferedBytes() + tr.presentStream.GetBufferedBytes()
}
// @Override
func (tr *TimestampColumnWriter) GetRetainedBytes() int64 {
retainedBytes := int64(TIMESTAMP_INSTANCE_SIZE) + tr.secondsStream.GetRetainedBytes() + tr.nanosStream.GetRetainedBytes() + tr.presentStream.GetRetainedBytes()
for _, statistics := range tr.rowGroupColumnStatistics.ToArray() {
retainedBytes += statistics.GetRetainedSizeInBytes()
}
return retainedBytes
}
// @Override
func (tr *TimestampColumnWriter) Reset() {
tr.closed = false
tr.secondsStream.Reset()
tr.nanosStream.Reset()
tr.presentStream.Reset()
tr.rowGroupColumnStatistics.Clear()
tr.statisticsBuilder = tr.statisticsBuilderSupplier.Get()
}
func (tr *TimestampColumnWriter) writeTimestampMicros(b block.Block) {
for i := util.INT32_ZERO; i < b.GetPositionCount(); i++ {
if !b.IsNull(i) {
micros := tr.kind.GetLong(b, i)
seconds := micros / int64(block.TTS_MICROSECONDS_PER_SECOND)
microsFraction := maths.FloorMod(micros, int64(block.TTS_MICROSECONDS_PER_SECOND))
nanosFraction := microsFraction * int64(block.TTS_NANOSECONDS_PER_MICROSECOND)
millis := maths.FloorDiv(micros, int64(block.TTS_MICROSECONDS_PER_MILLISECOND))
tr.writeValues(seconds, nanosFraction)
tr.statisticsBuilder.AddValue(millis)
}
}
}
func (tr *TimestampColumnWriter) writeTimestampNanos(b block.Block) {
for i := util.INT32_ZERO; i < b.GetPositionCount(); i++ {
if !b.IsNull(i) {
timestamp := tr.kind.GetObject(b, i).(*block.LongTimestamp)
seconds := timestamp.GetEpochMicros() / int64(block.TTS_MICROSECONDS_PER_SECOND)
microsFraction := maths.FloorMod(timestamp.GetEpochMicros(), int64(block.TTS_MICROSECONDS_PER_SECOND))
nanosFraction := (microsFraction * int64(block.TTS_NANOSECONDS_PER_MICROSECOND)) + int64(timestamp.GetPicosOfMicro()/block.TTS_PICOSECONDS_PER_NANOSECOND) // no rounding since the data has nanosecond precision, at most
millis := maths.FloorDiv(timestamp.GetEpochMicros(), int64(block.TTS_MICROSECONDS_PER_MILLISECOND))
tr.writeValues(seconds, nanosFraction)
tr.statisticsBuilder.AddValue(millis)
}
}
}
func (tr *TimestampColumnWriter) writeInstantShort(b block.Block) {
for i := util.INT32_ZERO; i < b.GetPositionCount(); i++ {
if !b.IsNull(i) {
tr.writeMillis(block.UnpackMillisUtc(tr.kind.GetLong(b, i)))
}
}
}
func (tr *TimestampColumnWriter) writeInstantLong(b block.Block) {
for i := util.INT32_ZERO; i < b.GetPositionCount(); i++ {
if !b.IsNull(i) {
timestamp := tr.kind.GetObject(b, i).(*block.LongTimestampWithTimeZone)
millis := timestamp.GetEpochMillis()
seconds := millis / int64(block.TTS_MILLISECONDS_PER_SECOND)
millisFraction := maths.FloorMod(millis, int64(block.TTS_MILLISECONDS_PER_SECOND))
nanosFraction := (millisFraction * int64(block.TTS_NANOSECONDS_PER_MILLISECOND)) + int64(timestamp.GetPicosOfMilli()/block.TTS_PICOSECONDS_PER_NANOSECOND)
tr.writeValues(seconds, nanosFraction)
tr.statisticsBuilder.AddValue(millis)
}
}
}
func (tr *TimestampColumnWriter) writeMillis(millis int64) {
seconds := millis / int64(block.TTS_MILLISECONDS_PER_SECOND)
millisFraction := maths.FloorMod(millis, int64(block.TTS_MILLISECONDS_PER_SECOND))
nanosFraction := millisFraction * int64(block.TTS_NANOSECONDS_PER_MILLISECOND)
tr.writeValues(seconds, nanosFraction)
tr.statisticsBuilder.AddValue(millis)
}
func (tr *TimestampColumnWriter) writeValues(seconds int64, nanosFraction int64) {
tr.secondsStream.WriteLong(seconds - TIMESTAMP_MOTH_EPOCH_IN_SECONDS)
tr.nanosStream.WriteLong(encodeNanos(nanosFraction))
}
func encodeNanos(nanos int64) int64 {
if nanos == 0 {
return 0
}
if (nanos % 100) != 0 {
return nanos << 3
}
nanos /= 100
trailingZeros := 1
for ((nanos % 10) == 0) && (trailingZeros < 7) {
nanos /= 10
trailingZeros++
}
return (nanos << 3) | int64(trailingZeros)
}