forked from pingcap/tidb
-
Notifications
You must be signed in to change notification settings - Fork 0
/
functions.go
382 lines (365 loc) · 11.3 KB
/
functions.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
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
// Copyright 2017 PingCAP, Inc.
//
// Licensed under the Apache License, Version 2.0 (the "License");
// you may not use this file except in compliance with the License.
// You may obtain a copy of the License at
//
// http://www.apache.org/licenses/LICENSE-2.0
//
// Unless required by applicable law or agreed to in writing, software
// distributed under the License is distributed on an "AS IS" BASIS,
// See the License for the specific language governing permissions and
// limitations under the License.
package json
import (
"bytes"
"encoding/binary"
"encoding/hex"
"fmt"
"unicode/utf8"
"github.com/juju/errors"
"github.com/pingcap/tidb/util/hack"
)
// Type returns type of JSON as string.
func (j JSON) Type() string {
switch j.TypeCode {
case TypeCodeObject:
return "OBJECT"
case TypeCodeArray:
return "ARRAY"
case TypeCodeLiteral:
switch byte(j.I64) {
case LiteralNil:
return "NULL"
default:
return "BOOLEAN"
}
case TypeCodeInt64:
return "INTEGER"
case TypeCodeUint64:
return "UNSIGNED INTEGER"
case TypeCodeFloat64:
return "DOUBLE"
case TypeCodeString:
return "STRING"
default:
msg := fmt.Sprintf(unknownTypeCodeErrorMsg, j.TypeCode)
panic(msg)
}
}
// Extract receives several path expressions as arguments, matches them in j, and returns:
// ret: target JSON matched any path expressions. maybe autowrapped as an array.
// found: true if any path expressions matched.
func (j JSON) Extract(pathExprList []PathExpression) (ret JSON, found bool) {
elemList := make([]JSON, 0, len(pathExprList))
for _, pathExpr := range pathExprList {
elemList = append(elemList, extract(j, pathExpr)...)
}
if len(elemList) == 0 {
found = false
} else if len(pathExprList) == 1 && len(elemList) == 1 {
// If pathExpr contains asterisks, len(elemList) won't be 1
// even if len(pathExprList) equals to 1.
found = true
ret = elemList[0]
} else {
found = true
ret.TypeCode = TypeCodeArray
ret.Array = append(ret.Array, elemList...)
}
return
}
// Unquote is for JSON_UNQUOTE.
func (j JSON) Unquote() (string, error) {
switch j.TypeCode {
case TypeCodeString:
s, err := unquoteString(j.Str)
if err != nil {
return "", errors.Trace(err)
}
// Remove prefix and suffix '"' and '\''.
slen := len(s)
if slen > 0 {
head, tail := s[0], s[slen-1]
if (head == '\'' && tail == '\'') || (head == '"' && tail == '"') {
return s[1 : slen-1], nil
}
}
return s, nil
default:
return j.String(), nil
}
}
// unquoteString recognizes the escape sequences shown in:
// https://dev.mysql.com/doc/refman/5.7/en/json-modification-functions.html#json-unquote-character-escape-sequences
func unquoteString(s string) (string, error) {
ret := new(bytes.Buffer)
for i := 0; i < len(s); i++ {
if s[i] == '\\' {
i++
if i == len(s) {
return "", errors.New("Missing a closing quotation mark in string")
}
switch s[i] {
case '"':
ret.WriteByte('"')
case 'b':
ret.WriteByte('\b')
case 'f':
ret.WriteByte('\f')
case 'n':
ret.WriteByte('\n')
case 'r':
ret.WriteByte('\r')
case 't':
ret.WriteByte('\t')
case '\\':
ret.WriteByte('\\')
case 'u':
if i+5 > len(s) {
return "", errors.Errorf("Invalid unicode: %s", s[i+1:])
}
char, size, err := decodeEscapedUnicode(hack.Slice(s[i+1 : i+5]))
if err != nil {
return "", errors.Trace(err)
}
ret.Write(char[0:size])
i += 5
default:
// For all other escape sequences, backslash is ignored.
ret.WriteByte(s[i])
}
} else {
ret.WriteByte(s[i])
}
}
return ret.String(), nil
}
// decodeEscapedUnicode decodes unicode into utf8 bytes specified in RFC 3629.
// According RFC 3629, the max length of utf8 characters is 4 bytes.
// And MySQL use 4 bytes to represent the unicode which must be in [0, 65536).
func decodeEscapedUnicode(s []byte) (char [4]byte, size int, err error) {
size, err = hex.Decode(char[0:2], s)
if err != nil || size != 2 {
// The unicode must can be represented in 2 bytes.
return char, 0, errors.Trace(err)
}
var unicode uint16
err = binary.Read(bytes.NewReader(char[0:2]), binary.BigEndian, &unicode)
if err != nil {
return char, 0, errors.Trace(err)
}
size = utf8.RuneLen(rune(unicode))
utf8.EncodeRune(char[0:size], rune(unicode))
return
}
// extract is used by Extract.
// NOTE: the return value will share something with j.
func extract(j JSON, pathExpr PathExpression) (ret []JSON) {
if len(pathExpr.legs) == 0 {
return []JSON{j}
}
currentLeg, subPathExpr := pathExpr.popOneLeg()
if currentLeg.typ == pathLegIndex {
// If j is not an array, autowrap that into array.
if j.TypeCode != TypeCodeArray {
j = autoWrapAsArray(j, 1)
}
if currentLeg.arrayIndex == arrayIndexAsterisk {
for _, child := range j.Array {
ret = append(ret, extract(child, subPathExpr)...)
}
} else if currentLeg.arrayIndex < len(j.Array) {
childRet := extract(j.Array[currentLeg.arrayIndex], subPathExpr)
ret = append(ret, childRet...)
}
} else if currentLeg.typ == pathLegKey && j.TypeCode == TypeCodeObject {
if len(currentLeg.dotKey) == 1 && currentLeg.dotKey[0] == '*' {
var sortedKeys = getSortedKeys(j.Object) // iterate over sorted keys.
for _, child := range sortedKeys {
ret = append(ret, extract(j.Object[child], subPathExpr)...)
}
} else if child, ok := j.Object[currentLeg.dotKey]; ok {
childRet := extract(child, subPathExpr)
ret = append(ret, childRet...)
}
} else if currentLeg.typ == pathLegDoubleAsterisk {
ret = append(ret, extract(j, subPathExpr)...)
if j.TypeCode == TypeCodeArray {
for _, child := range j.Array {
ret = append(ret, extract(child, pathExpr)...)
}
} else if j.TypeCode == TypeCodeObject {
var sortedKeys = getSortedKeys(j.Object)
for _, child := range sortedKeys {
ret = append(ret, extract(j.Object[child], pathExpr)...)
}
}
}
return
}
// autoWrapAsArray wraps input JSON into an array if needed.
func autoWrapAsArray(j JSON, hintLength int) JSON {
jnew := CreateJSON(nil)
jnew.TypeCode = TypeCodeArray
jnew.Array = make([]JSON, 0, hintLength)
jnew.Array = append(jnew.Array, j)
return jnew
}
// Merge merges suffixes into j according the following rules:
// 1) adjacent arrays are merged to a single array;
// 2) adjacent object are merged to a single object;
// 3) a scalar value is autowrapped as an array before merge;
// 4) an adjacent array and object are merged by autowrapping the object as an array.
func (j JSON) Merge(suffixes []JSON) JSON {
if j.TypeCode != TypeCodeArray && j.TypeCode != TypeCodeObject {
j = autoWrapAsArray(j, len(suffixes)+1)
}
for i := 0; i < len(suffixes); i++ {
suffix := suffixes[i]
switch j.TypeCode {
case TypeCodeArray:
if suffix.TypeCode == TypeCodeArray {
// rule (1)
j.Array = append(j.Array, suffix.Array...)
} else {
// rule (3), (4)
j.Array = append(j.Array, suffix)
}
case TypeCodeObject:
if suffix.TypeCode == TypeCodeObject {
// rule (2)
for key := range suffix.Object {
if child, ok := j.Object[key]; ok {
j.Object[key] = child.Merge([]JSON{suffix.Object[key]})
} else {
j.Object[key] = suffix.Object[key]
}
}
} else {
// rule (4)
j = autoWrapAsArray(j, len(suffixes)+1-i)
if suffix.TypeCode == TypeCodeArray {
j.Array = append(j.Array, suffix.Array...)
} else {
j.Array = append(j.Array, suffix)
}
}
}
}
return j
}
// ModifyType is for modify a JSON. There are three valid values:
// ModifyInsert, ModifyReplace and ModifySet.
type ModifyType byte
const (
// ModifyInsert is for insert a new element into a JSON.
ModifyInsert ModifyType = 0x01
// ModifyReplace is for replace an old elemList from a JSON.
ModifyReplace ModifyType = 0x02
// ModifySet = ModifyInsert | ModifyReplace
ModifySet ModifyType = 0x03
)
// Modify modifies a JSON object by insert, replace or set.
// All path expressions cannot contain * or ** wildcard.
// If any error occurs, the input won't be changed.
func (j JSON) Modify(pathExprList []PathExpression, values []JSON, mt ModifyType) (retj JSON, err error) {
if len(pathExprList) != len(values) {
// TODO: should return 1582(42000)
return retj, errors.New("Incorrect parameter count")
}
for _, pathExpr := range pathExprList {
if pathExpr.flags.containsAnyAsterisk() {
// TODO: should return 3149(42000)
return retj, errors.New("Invalid path expression")
}
}
for i := 0; i < len(pathExprList); i++ {
pathExpr, value := pathExprList[i], values[i]
j = set(j, pathExpr, value, mt)
}
return j, nil
}
// set is for Modify. The result JSON maybe share something with input JSON.
func set(j JSON, pathExpr PathExpression, value JSON, mt ModifyType) JSON {
if len(pathExpr.legs) == 0 {
if mt&ModifyReplace != 0 {
return value
}
return j
}
currentLeg, subPathExpr := pathExpr.popOneLeg()
if currentLeg.typ == pathLegIndex {
// If j is not an array, we should autowrap that as array.
// Then if its length equals to 1, we unwrap it back.
var shouldUnwrap = false
if j.TypeCode != TypeCodeArray {
j = autoWrapAsArray(j, 1)
shouldUnwrap = true
}
var index = currentLeg.arrayIndex
if len(j.Array) > index {
// e.g. json_replace('[1, 2, 3]', '$[0]', "x") => '["x", 2, 3]'
j.Array[index] = set(j.Array[index], subPathExpr, value, mt)
} else if len(subPathExpr.legs) == 0 && mt&ModifyInsert != 0 {
// e.g. json_insert('[1, 2, 3]', '$[3]', "x") => '[1, 2, 3, "x"]'
j.Array = append(j.Array, value)
}
if len(j.Array) == 1 && shouldUnwrap {
j = j.Array[0]
}
} else if currentLeg.typ == pathLegKey && j.TypeCode == TypeCodeObject {
var key = currentLeg.dotKey
if child, ok := j.Object[key]; ok {
// e.g. json_replace('{"a": 1}', '$.a', 2) => '{"a": 2}'
j.Object[key] = set(child, subPathExpr, value, mt)
} else if len(subPathExpr.legs) == 0 && mt&ModifyInsert != 0 {
// e.g. json_insert('{"a": 1}', '$.b', 2) => '{"a": 1, "b": 2}'
j.Object[key] = value
}
}
// For these cases, we just return the input JSON back without any change:
// 1) we want to insert a new element, but the full path has already exists;
// 2) we want to replace an old element, but the full path doesn't exist;
// 3) we want to insert or replace something, but the path without last leg doesn't exist.
return j
}
// Remove removes the elements indicated by pathExprList from JSON.
func (j JSON) Remove(pathExprList []PathExpression) (JSON, error) {
for _, pathExpr := range pathExprList {
if len(pathExpr.legs) == 0 {
// TODO: should return 3153(42000)
return j, errors.New("Invalid path expression")
}
if pathExpr.flags.containsAnyAsterisk() {
// TODO: should return 3149(42000)
return j, errors.New("Invalid path expression")
}
j = remove(j, pathExpr)
}
return j, nil
}
// remove is used in Remove.
func remove(j JSON, pathExpr PathExpression) JSON {
currentLeg, subPathExpr := pathExpr.popOneLeg()
if currentLeg.typ == pathLegIndex && j.TypeCode == TypeCodeArray {
var index = currentLeg.arrayIndex
if len(j.Array) > index {
if len(subPathExpr.legs) == 0 {
j.Array = append(j.Array[0:index], j.Array[index+1:]...)
} else {
j.Array[index] = remove(j.Array[index], subPathExpr)
}
}
} else if currentLeg.typ == pathLegKey && j.TypeCode == TypeCodeObject {
var key = currentLeg.dotKey
if child, ok := j.Object[key]; ok {
if len(subPathExpr.legs) == 0 {
delete(j.Object, key)
} else {
j.Object[key] = remove(child, subPathExpr)
}
}
}
return j
}