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
|
package dataset
import (
"context"
"errors"
"fmt"
"iter"
"path/filepath"
"slices"
"github.com/apache/arrow-go/v18/arrow"
"github.com/apache/arrow-go/v18/arrow/array"
"github.com/apache/arrow-go/v18/arrow/memory"
"github.com/apache/arrow-go/v18/parquet/file"
"github.com/apache/arrow-go/v18/parquet/pqarrow"
)
type ParquetReader struct {
shards []string
textColumn string
batchSize int64
err error
}
func NewParquetReader(name string) (*ParquetReader, error) {
var shards []string
if matches, err := filepath.Glob(filepath.Join(name, "*.parquet")); err != nil {
return nil, err
} else {
shards = matches
}
if len(shards) == 0 {
return nil, fmt.Errorf("no parquet files in %q", name)
}
slices.Sort(shards)
r := &ParquetReader{
shards: shards,
textColumn: "text",
batchSize: 1024,
}
return r, nil
}
func (p *ParquetReader) Num() (int, error) {
r := 0
for _, name := range p.shards {
n, err := numRows(name)
if err != nil {
return 0, err
}
r += int(n)
}
return r, nil
}
func (p *ParquetReader) Err() error {
return p.err
}
func (p *ParquetReader) Texts() iter.Seq2[int, string] {
return func(yield func(int, string) bool) {
n := 0
for _, name := range p.shards {
for text := range p.read(name) {
if p.err != nil {
return
}
if !yield(n, text) {
return
}
n++
}
}
}
}
func (p *ParquetReader) read(name string) iter.Seq[string] {
return func(yield func(string) bool) {
var reader *file.Reader
if r, err := file.OpenParquetFile(name, false); err != nil {
p.err = err
return
} else {
reader = r
}
defer reader.Close()
var arrowReader *pqarrow.FileReader
if r, err := pqarrow.NewFileReader(reader, pqarrow.ArrowReadProperties{BatchSize: p.batchSize}, memory.DefaultAllocator); err != nil {
p.err = err
return
} else {
arrowReader = r
}
var schema *arrow.Schema
if s, err := arrowReader.Schema(); err != nil {
p.err = err
return
} else {
schema = s
}
idxs := schema.FieldIndices(p.textColumn)
if len(idxs) == 0 {
p.err = errors.New("unknown column")
return
}
var recordReader pqarrow.RecordReader
if r, err := arrowReader.GetRecordReader(context.TODO(), []int{idxs[0]}, nil); err != nil {
p.err = err
return
} else {
recordReader = r
}
defer recordReader.Release()
for recordReader.Next() {
record := recordReader.RecordBatch()
text, ok := record.Column(0).(*array.String)
if !ok {
p.err = fmt.Errorf("unexpected column type")
return
}
for i := range int(record.NumRows()) {
if !text.IsValid(i) {
continue
}
if !yield(text.Value(i)) {
return
}
}
}
}
}
func numRows(name string) (int64, error) {
var reader *file.Reader
if r, err := file.OpenParquetFile(name, false); err != nil {
return 0, err
} else {
reader = r
defer reader.Close()
}
return reader.NumRows(), nil
}
|
