1
|
|
|
package csv |
2
|
|
|
|
3
|
|
|
import ( |
4
|
|
|
"adrianolaselva.github.io/csvql/pkg/filehandler" |
5
|
|
|
"adrianolaselva.github.io/csvql/pkg/storage" |
6
|
|
|
"bytes" |
7
|
|
|
"database/sql" |
8
|
|
|
"encoding/csv" |
9
|
|
|
"fmt" |
10
|
|
|
"github.com/schollz/progressbar/v3" |
11
|
|
|
"io" |
12
|
|
|
"os" |
13
|
|
|
) |
14
|
|
|
|
15
|
|
|
type csvHandler struct { |
16
|
|
|
bar *progressbar.ProgressBar |
17
|
|
|
storage storage.Storage |
18
|
|
|
file *os.File |
19
|
|
|
fileInput string |
20
|
|
|
lines int |
21
|
|
|
delimiter rune |
22
|
|
|
} |
23
|
|
|
|
24
|
|
|
func NewCsvHandler(fileInput string, delimiter rune, bar *progressbar.ProgressBar, storage storage.Storage) filehandler.FileHandler { |
25
|
|
|
return &csvHandler{fileInput: fileInput, delimiter: delimiter, storage: storage, bar: bar} |
26
|
|
|
} |
27
|
|
|
|
28
|
|
|
// Import import data |
29
|
|
|
func (c *csvHandler) Import() error { |
30
|
|
|
if err := c.openFile(); err != nil { |
31
|
|
|
return err |
32
|
|
|
} |
33
|
|
|
|
34
|
|
|
if err := c.loadTotalRows(); err != nil { |
35
|
|
|
return err |
36
|
|
|
} |
37
|
|
|
|
38
|
|
|
if err := c.loadDataFromFile(); err != nil { |
39
|
|
|
return err |
40
|
|
|
} |
41
|
|
|
|
42
|
|
|
return nil |
43
|
|
|
} |
44
|
|
|
|
45
|
|
|
// Query execute statements |
46
|
|
|
func (c *csvHandler) Query(cmd string) (*sql.Rows, error) { |
47
|
|
|
return c.storage.Query(cmd) |
48
|
|
|
} |
49
|
|
|
|
50
|
|
|
// Lines return total lines |
51
|
|
|
func (c *csvHandler) Lines() int { |
52
|
|
|
return c.lines |
53
|
|
|
} |
54
|
|
|
|
55
|
|
|
// Close execute in defer |
56
|
|
|
func (c *csvHandler) Close() error { |
57
|
|
|
defer func(storage storage.Storage) { |
58
|
|
|
_ = storage.Close() |
59
|
|
|
}(c.storage) |
60
|
|
|
|
61
|
|
|
defer func(file *os.File) { |
62
|
|
|
_ = file.Close() |
63
|
|
|
}(c.file) |
64
|
|
|
|
65
|
|
|
return nil |
66
|
|
|
} |
67
|
|
|
|
68
|
|
|
func (c *csvHandler) loadDataFromFile() error { |
69
|
|
|
c.bar.ChangeMax(c.lines) |
70
|
|
|
|
71
|
|
|
r := csv.NewReader(c.file) |
72
|
|
|
r.Comma = c.delimiter |
73
|
|
|
|
74
|
|
|
headers, err := r.Read() |
75
|
|
|
if err != nil { |
76
|
|
|
return fmt.Errorf("failed to load headers: %s", err) |
77
|
|
|
} |
78
|
|
|
|
79
|
|
|
if err := c.storage.SetColumns(headers).BuildStructure(); err != nil { |
80
|
|
|
return fmt.Errorf("failed to load headers and build structure: %s", err) |
81
|
|
|
} |
82
|
|
|
|
83
|
|
|
line := 1 |
84
|
|
|
for { |
85
|
|
|
line++ |
86
|
|
|
records, err := r.Read() |
87
|
|
|
if err == io.EOF { |
88
|
|
|
break |
89
|
|
|
} |
90
|
|
|
|
91
|
|
|
var values []any |
92
|
|
|
for _, r := range records { |
93
|
|
|
values = append(values, r) |
94
|
|
|
} |
95
|
|
|
|
96
|
|
|
_ = c.bar.Add(1) |
97
|
|
|
if err := c.storage.InsertRow(values); err != nil { |
98
|
|
|
return fmt.Errorf("failed to process row: %s", err) |
99
|
|
|
} |
100
|
|
|
} |
101
|
|
|
|
102
|
|
|
return nil |
103
|
|
|
} |
104
|
|
|
|
105
|
|
|
// openFile open file |
106
|
|
|
func (c *csvHandler) openFile() error { |
107
|
|
|
f, err := os.Open(c.fileInput) |
108
|
|
|
if err != nil { |
109
|
|
|
return fmt.Errorf("failed to open file: %s", err) |
110
|
|
|
} |
111
|
|
|
|
112
|
|
|
c.file = f |
113
|
|
|
|
114
|
|
|
return nil |
115
|
|
|
} |
116
|
|
|
|
117
|
|
|
// loadTotalRows load total rows in file |
118
|
|
|
func (c *csvHandler) loadTotalRows() error { |
119
|
|
|
r, err := os.Open(c.fileInput) |
120
|
|
|
if err != nil { |
121
|
|
|
return err |
122
|
|
|
} |
123
|
|
|
defer r.Close() |
124
|
|
|
|
125
|
|
|
buf := make([]byte, 32*1024) |
126
|
|
|
c.lines = 0 |
127
|
|
|
lineSep := []byte{'\n'} |
128
|
|
|
|
129
|
|
|
for { |
130
|
|
|
r, err := r.Read(buf) |
131
|
|
|
c.lines += bytes.Count(buf[:r], lineSep) |
132
|
|
|
|
133
|
|
|
switch { |
134
|
|
|
case err == io.EOF: |
135
|
|
|
return nil |
136
|
|
|
|
137
|
|
|
case err != nil: |
138
|
|
|
return fmt.Errorf("failed to totalize rows: %s", err) |
139
|
|
|
} |
140
|
|
|
} |
141
|
|
|
} |
142
|
|
|
|