1
|
|
|
<?php |
2
|
|
|
|
3
|
|
|
/* |
4
|
|
|
* This file is part of YaEtl. |
5
|
|
|
* (c) Fabrice de Stefanis / https://github.com/fab2s/YaEtl |
6
|
|
|
* This source file is licensed under the MIT license which you will |
7
|
|
|
* find in the LICENSE file or at https://opensource.org/licenses/MIT |
8
|
|
|
*/ |
9
|
|
|
|
10
|
|
|
namespace fab2s\YaEtl\Extractors; |
11
|
|
|
|
12
|
|
|
/** |
13
|
|
|
* Abstract Class UniqueKeyExtractorAbstract |
14
|
|
|
*/ |
15
|
|
|
abstract class UniqueKeyExtractorAbstract extends DbExtractorAbstract implements JoinableInterface |
16
|
|
|
{ |
17
|
|
|
/** |
18
|
|
|
* the unique key name |
19
|
|
|
* |
20
|
|
|
* @var array|string |
21
|
|
|
*/ |
22
|
|
|
protected $compositeKey; |
23
|
|
|
|
24
|
|
|
/** |
25
|
|
|
* @var string |
26
|
|
|
*/ |
27
|
|
|
protected $uniqueKeyName; |
28
|
|
|
|
29
|
|
|
/** |
30
|
|
|
* @var string |
31
|
|
|
*/ |
32
|
|
|
protected $uniqueKeyAlias; |
33
|
|
|
|
34
|
|
|
/** |
35
|
|
|
* @var array |
36
|
|
|
*/ |
37
|
|
|
protected $uniqueKeyValues = []; |
38
|
|
|
|
39
|
|
|
/** |
40
|
|
|
* @var array |
41
|
|
|
*/ |
42
|
|
|
protected $uniqueKeyValueBuffer = []; |
43
|
|
|
|
44
|
|
|
/** |
45
|
|
|
* @var callable |
46
|
|
|
*/ |
47
|
|
|
protected $merger; |
48
|
|
|
|
49
|
|
|
/** |
50
|
|
|
* @var OnClauseInterface |
51
|
|
|
*/ |
52
|
|
|
protected $onClose; |
53
|
|
|
|
54
|
|
|
/** |
55
|
|
|
* @var array of OnClauseInterface |
56
|
|
|
*/ |
57
|
|
|
protected $joinerOnCloses = []; |
58
|
|
|
|
59
|
|
|
/** |
60
|
|
|
* the record map |
61
|
|
|
* |
62
|
|
|
* @var array |
63
|
|
|
*/ |
64
|
|
|
protected $recordMap; |
65
|
|
|
|
66
|
|
|
/** |
67
|
|
|
* The Joignable we may be joining against |
68
|
|
|
* |
69
|
|
|
* @var JoignableInterface |
70
|
|
|
*/ |
71
|
|
|
protected $joinFrom; |
72
|
|
|
|
73
|
|
|
/** |
74
|
|
|
* generic extraction from tables with unique (composite) key |
75
|
|
|
* |
76
|
|
|
* @param string $extractQuery |
77
|
|
|
* @param array|string $uniqueKey can be either a unique key name as |
78
|
|
|
* string |
79
|
|
|
* '(table.)compositeKeyName' // ('id' by default) |
80
|
|
|
* |
81
|
|
|
* or an array : |
82
|
|
|
* ['(table.)compositeKey1'] // single unique key |
83
|
|
|
* ['(table.)compositeKey1', '(table.)compositeKey2', ] // composite unique key |
84
|
|
|
* |
85
|
|
|
* or an associative array in case you are using aliases : |
86
|
|
|
* [ |
87
|
|
|
* '(table.)compositeKey1' => 'aliasNameAsInRecord', |
88
|
|
|
* ] |
89
|
|
|
* |
90
|
|
|
* and : |
91
|
|
|
* [ |
92
|
|
|
* '(table.)compositeKey1' => 'aliasNameAsInRecord1', |
93
|
|
|
* '(table.)compositeKey2' => 'aliasNameAsInRecord2', |
94
|
|
|
* // ... |
95
|
|
|
* ] |
96
|
|
|
*/ |
97
|
|
|
public function __construct($extractQuery = null, $uniqueKey = 'id') |
98
|
|
|
{ |
99
|
|
|
$this->configureUniqueKey($uniqueKey); |
100
|
|
|
|
101
|
|
|
parent::__construct($extractQuery); |
102
|
|
|
} |
103
|
|
|
|
104
|
|
|
/** |
105
|
|
|
* get Joiner's ON clause. Only used in Join mode |
106
|
|
|
* |
107
|
|
|
* @return OnClauseInterface |
108
|
|
|
*/ |
109
|
|
|
public function getOnClause() |
110
|
|
|
{ |
111
|
|
|
return $this->onClose; |
112
|
|
|
} |
113
|
|
|
|
114
|
|
|
/** |
115
|
|
|
* Set Joiner's ON clause. Only used in Join mode |
116
|
|
|
* |
117
|
|
|
* @param OnClauseInterface $onClause |
118
|
|
|
* |
119
|
|
|
* @return $this |
120
|
|
|
*/ |
121
|
|
|
public function setOnClause(OnClauseInterface $onClause) |
122
|
|
|
{ |
123
|
|
|
$this->onClose = $onClause; |
124
|
|
|
|
125
|
|
|
return $this; |
126
|
|
|
} |
127
|
|
|
|
128
|
|
|
/** |
129
|
|
|
* register ON clause field mapping. Used by an eventual joiner to this |
130
|
|
|
* |
131
|
|
|
* @param OnClauseInterface $onClause |
132
|
|
|
* |
133
|
|
|
* @return $this |
134
|
|
|
*/ |
135
|
|
|
public function registerJoinerOnClause(OnClauseInterface $onClause) |
136
|
|
|
{ |
137
|
|
|
$this->joinerOnCloses[] = $onClause; |
138
|
|
|
|
139
|
|
|
return $this; |
140
|
|
|
} |
141
|
|
|
|
142
|
|
|
/** |
143
|
|
|
* @param JoinableInterface $joinFrom |
144
|
|
|
* |
145
|
|
|
* @throws \Exception |
146
|
|
|
* |
147
|
|
|
* @return $this |
148
|
|
|
*/ |
149
|
|
|
public function setJoinFrom(JoinableInterface $joinFrom) |
150
|
|
|
{ |
151
|
|
|
if (preg_match('`^(.+)(order\s+by.*)$`is', $this->extractQuery)) { |
152
|
|
|
throw new \Exception("[YaEtl] A Joiner must not order its query got: $this->extractQuery"); |
153
|
|
|
} |
154
|
|
|
|
155
|
|
|
// since we are joining, we are not a traversable anymore |
156
|
|
|
$this->isATraversable = false; |
157
|
|
|
// and we return a value |
158
|
|
|
$this->isAReturningVal = true; |
159
|
|
|
$this->joinFrom = $joinFrom; |
|
|
|
|
160
|
|
|
|
161
|
|
|
return $this; |
162
|
|
|
} |
163
|
|
|
|
164
|
|
|
/** |
165
|
|
|
* @param string $fromKeyAlias The from unique key to get the map against |
166
|
|
|
* as exposed in the record |
167
|
|
|
* |
168
|
|
|
* @return array [keyValue1, keyValue2, ...] |
169
|
|
|
*/ |
170
|
|
|
public function getRecordMap($fromKeyAlias = null) |
171
|
|
|
{ |
172
|
|
|
return $fromKeyAlias === null ? $this->recordMap : $this->recordMap[$fromKeyAlias]; |
173
|
|
|
} |
174
|
|
|
|
175
|
|
|
/** |
176
|
|
|
* @param mixed $param |
177
|
|
|
* |
178
|
|
|
* @return bool |
179
|
|
|
*/ |
180
|
|
|
public function extract($param = null) |
181
|
|
|
{ |
182
|
|
|
if (isset($this->joinFrom)) { |
183
|
|
|
// join mode, get record map |
184
|
|
|
$this->uniqueKeyValues = $this->joinFrom->getRecordMap($this->onClose->getFromKeyAlias()); |
185
|
|
|
// limit does not apply in join mode |
186
|
|
|
$this->enforceBatchSize(); |
187
|
|
|
if (empty($this->uniqueKeyValues)) { |
188
|
|
|
return false; |
189
|
|
|
} |
190
|
|
|
|
191
|
|
|
if ($this->fetchRecords()) { |
192
|
|
|
// gen record map before we set defaults |
193
|
|
|
$this->genRecordMap() |
194
|
|
|
->setDefaultExtracted(); |
195
|
|
|
|
196
|
|
|
return true; |
197
|
|
|
} |
198
|
|
|
|
199
|
|
|
return false; |
200
|
|
|
} |
201
|
|
|
|
202
|
|
|
// enforce limit if any is set |
203
|
|
|
if ($this->isLimitReached()) { |
204
|
|
|
return false; |
205
|
|
|
} |
206
|
|
|
|
207
|
|
|
$this->enforceBatchSize(); |
208
|
|
|
if ($this->fetchRecords()) { |
209
|
|
|
$this->incrementOffset() |
210
|
|
|
->genRecordMap(); |
211
|
|
|
|
212
|
|
|
return true; |
213
|
|
|
} |
214
|
|
|
|
215
|
|
|
return false; |
216
|
|
|
} |
217
|
|
|
|
218
|
|
|
/** |
219
|
|
|
* @return $this |
220
|
|
|
*/ |
221
|
|
|
public function enforceBatchSize() |
222
|
|
|
{ |
223
|
|
|
if (isset($this->joinFrom)) { |
224
|
|
|
// obey batch size to allow fromer to fetch a huge amount of records |
225
|
|
|
// while keeping the where in query size under control by splitting |
226
|
|
|
// it into several chunks. |
227
|
|
|
if (!empty($this->uniqueKeyValueBuffer)) { |
228
|
|
|
// happend eventual new keys from upstream extractor |
229
|
|
|
\end($this->uniqueKeyValueBuffer); |
230
|
|
|
$lastKeyValue = \current($this->uniqueKeyValueBuffer); |
231
|
|
|
\reset($this->uniqueKeyValueBuffer); |
232
|
|
|
$set = false; |
233
|
|
|
foreach ($this->uniqueKeyValues as $uniqueKeyValue) { |
234
|
|
|
// move forward in the array until we reach the |
235
|
|
|
// last value we already had in the buffer and |
236
|
|
|
// then happend the rest |
237
|
|
|
if (!$set && $uniqueKeyValue === $lastKeyValue) { |
238
|
|
|
$set = true; |
239
|
|
|
continue; |
240
|
|
|
} |
241
|
|
|
|
242
|
|
|
$this->uniqueKeyValueBuffer[$uniqueKeyValue] = $uniqueKeyValue; |
243
|
|
|
} |
244
|
|
|
} else { |
245
|
|
|
// get everything |
246
|
|
|
$this->uniqueKeyValueBuffer = $this->uniqueKeyValues; |
247
|
|
|
} |
248
|
|
|
|
249
|
|
|
$this->uniqueKeyValues = \array_slice($this->uniqueKeyValueBuffer, 0, $this->batchSize, true); |
250
|
|
|
$this->uniqueKeyValueBuffer = \array_slice($this->uniqueKeyValueBuffer, $this->batchSize, null, true); |
251
|
|
|
|
252
|
|
|
return $this; |
253
|
|
|
} |
254
|
|
|
|
255
|
|
|
parent::enforceBatchSize(); |
256
|
|
|
|
257
|
|
|
return $this; |
258
|
|
|
} |
259
|
|
|
|
260
|
|
|
/** |
261
|
|
|
* @param mixed $record |
262
|
|
|
* |
263
|
|
|
* @return mixed The result of the join |
264
|
|
|
*/ |
265
|
|
|
public function exec($record) |
266
|
|
|
{ |
267
|
|
|
$uniqueKeyValue = $record[$this->uniqueKeyAlias]; |
268
|
|
|
|
269
|
|
|
if (isset($this->extracted[$uniqueKeyValue])) { |
270
|
|
|
$joinRecord = $this->extracted[$uniqueKeyValue]; |
271
|
|
|
unset($this->extracted[$uniqueKeyValue]); |
272
|
|
|
if ($joinRecord === false) { |
273
|
|
|
// skip record |
274
|
|
|
$this->carrier->continueFlow(); |
275
|
|
|
|
276
|
|
|
return $record; |
277
|
|
|
} |
278
|
|
|
|
279
|
|
|
++$this->numRecords; |
280
|
|
|
|
281
|
|
|
return $this->onClose->merge($record, $joinRecord); |
282
|
|
|
} elseif (isset($this->uniqueKeyValueBuffer[$uniqueKeyValue])) { |
283
|
|
|
// uniqueKeyValueBuffer should never run out until |
284
|
|
|
// the fromer stop providing records which mean |
285
|
|
|
// we do not reach here. Case left is when this |
286
|
|
|
// batchSize < fromer batchSize. |
287
|
|
|
// trigger extract |
288
|
|
|
if ($this->extract()) { |
289
|
|
|
return $this->exec($record); |
290
|
|
|
} |
291
|
|
|
} elseif ($this->extract()) { |
292
|
|
|
return $this->exec($record); |
293
|
|
|
} |
294
|
|
|
|
295
|
|
|
// something is wrong |
296
|
|
|
throw new \Exception('[YaEtl] Record map missmatch betwen Joiner ' . \get_class($this) . ' and Fromer ' . \get_class($this->joinFrom)); |
297
|
|
|
} |
298
|
|
|
|
299
|
|
|
/** |
300
|
|
|
* @param array|string $uniqueKey can be either a unique key name as |
301
|
|
|
* string |
302
|
|
|
* '(table.)compositeKeyName' // ('id' by default) |
303
|
|
|
* |
304
|
|
|
* or an array : |
305
|
|
|
* ['(table.)compositeKey1'] // single unique key |
306
|
|
|
* ['(table.)compositeKey1', '(table.)compositeKey2', ] // composite unique key |
307
|
|
|
* |
308
|
|
|
* or an associative array in case you are using aliases : |
309
|
|
|
* [ |
310
|
|
|
* '(table.)compositeKey1' => 'aliasNameAsInRecord', |
311
|
|
|
* ] |
312
|
|
|
* |
313
|
|
|
* and : |
314
|
|
|
* [ |
315
|
|
|
* '(table.)compositeKey1' => 'aliasNameAsInRecord1', |
316
|
|
|
* '(table.)compositeKey2' => 'aliasNameAsInRecord2', |
317
|
|
|
* // ... |
318
|
|
|
* ] |
319
|
|
|
* |
320
|
|
|
* @return $this |
321
|
|
|
*/ |
322
|
|
|
protected function configureUniqueKey($uniqueKey) |
323
|
|
|
{ |
324
|
|
|
$uniqueKey = \is_array($uniqueKey) ? $uniqueKey : [$uniqueKey]; |
325
|
|
|
$this->compositeKey = []; |
326
|
|
|
$this->uniqueKeyName = null; |
327
|
|
|
$this->uniqueKeyAlias = null; |
328
|
|
|
foreach ($uniqueKey as $key => $value) { |
329
|
|
|
if (\is_numeric($key)) { |
330
|
|
|
$compositeKeyName = $this->cleanUpKeyName($value); |
331
|
|
|
$compositeKeyParts = \explode('.', $compositeKeyName); |
332
|
|
|
$compositeKeyAlias = \end($compositeKeyParts); |
333
|
|
|
} else { |
334
|
|
|
$compositeKeyName = $this->cleanUpKeyName($key); |
335
|
|
|
$compositeKeyAlias = $this->cleanUpKeyName($value); |
336
|
|
|
} |
337
|
|
|
|
338
|
|
|
$this->compositeKey[$compositeKeyName] = $compositeKeyAlias; |
339
|
|
|
} |
340
|
|
|
|
341
|
|
|
if (\count($this->compositeKey) === 1) { |
342
|
|
|
$this->uniqueKeyName = \key($this->compositeKey); |
343
|
|
|
$this->uniqueKeyAlias = \current($this->compositeKey); |
344
|
|
|
} |
345
|
|
|
|
346
|
|
|
return $this; |
347
|
|
|
} |
348
|
|
|
|
349
|
|
|
/** |
350
|
|
|
* @param string $keyName |
351
|
|
|
* |
352
|
|
|
* @return string |
353
|
|
|
*/ |
354
|
|
|
protected function cleanUpKeyName($keyName) |
355
|
|
|
{ |
356
|
|
|
return \trim($keyName, '` '); |
357
|
|
|
} |
358
|
|
|
|
359
|
|
|
/** |
360
|
|
|
* prepare record set to obey join mode eg return record = true |
361
|
|
|
* to break branch execution when no match are found in join more |
362
|
|
|
* or default to be later merged in left join mode |
363
|
|
|
* |
364
|
|
|
* @return $this |
365
|
|
|
*/ |
366
|
|
|
protected function setDefaultExtracted() |
367
|
|
|
{ |
368
|
|
|
if ($this->joinFrom !== null) { |
369
|
|
|
$defaultExtracted = \array_fill_keys($this->uniqueKeyValues, $this->onClose->isLeftJoin() ? $this->onClose->getDefaultRecord() : false); |
370
|
|
|
|
371
|
|
|
foreach ($defaultExtracted as $keyValue => &$default) { |
372
|
|
|
// extracted is an array in join mode |
373
|
|
|
if (isset($this->extracted[$keyValue])) { |
374
|
|
|
$default = $this->extracted[$keyValue]; |
375
|
|
|
} |
376
|
|
|
} |
377
|
|
|
|
378
|
|
|
$this->extracted = $defaultExtracted; |
379
|
|
|
} |
380
|
|
|
|
381
|
|
|
return $this; |
382
|
|
|
} |
383
|
|
|
|
384
|
|
|
/** |
385
|
|
|
* @return $this |
386
|
|
|
*/ |
387
|
|
|
protected function genRecordMap() |
388
|
|
|
{ |
389
|
|
|
// here we need to build record map ready for all joiners |
390
|
|
|
$this->recordMap = []; |
391
|
|
|
foreach ($this->joinerOnCloses as $onClose) { |
392
|
|
|
$fromKeyAlias = $onClose->getFromKeyAlias(); |
393
|
|
|
if (isset($this->recordMap[$fromKeyAlias])) { |
394
|
|
|
// looks like there is more than |
395
|
|
|
// one joiner on this key |
396
|
|
|
continue; |
397
|
|
|
} |
398
|
|
|
|
399
|
|
|
// generate rercord map |
400
|
|
|
$this->recordMap[$fromKeyAlias] = []; |
401
|
|
|
|
402
|
|
|
$map = &$this->recordMap[$fromKeyAlias]; |
403
|
|
|
// we do not want to map defaults here as we do not want joiners |
404
|
|
|
// to this to join on null |
405
|
|
|
// we could optimize a little bit for cases where |
406
|
|
|
// $this->extracted is an indexed array on the proper key but ... |
407
|
|
|
foreach ($this->extracted as $record) { |
408
|
|
|
if (!isset($record[$fromKeyAlias])) { |
409
|
|
|
// Since we do not enforce key alias existance during init |
410
|
|
|
// we have to do it here |
411
|
|
|
throw new \Exception("[YaEtl] From Key Alias not found in record: $fromKeyAlias"); |
412
|
|
|
} |
413
|
|
|
|
414
|
|
|
$fromKeyValue = $record[$fromKeyAlias]; |
415
|
|
|
$map[$fromKeyValue] = $fromKeyValue; |
416
|
|
|
} |
417
|
|
|
} |
418
|
|
|
|
419
|
|
|
return $this; |
420
|
|
|
} |
421
|
|
|
} |
422
|
|
|
|
Our type inference engine has found an assignment to a property that is incompatible with the declared type of that property.
Either this assignment is in error or the assigned type should be added to the documentation/type hint for that property..