-
Notifications
You must be signed in to change notification settings - Fork 2
/
Medley.ecl
1350 lines (1197 loc) · 58.6 KB
/
Medley.ecl
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
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
938
939
940
941
942
943
944
945
946
947
948
949
950
951
952
953
954
955
956
957
958
959
960
961
962
963
964
965
966
967
968
969
970
971
972
973
974
975
976
977
978
979
980
981
982
983
984
985
986
987
988
989
990
991
992
993
994
995
996
997
998
999
1000
/**
* Code for performing record fragmentation matching. The module provides
* the following functionality:
*
* - Creation of search indexes, linked back to entity ID values
* - Fuzzy matching within the source dataset, linking similar IDs
* - Searching for related IDs, given a dataset of IDs
* - Searching for related IDs, given a dataset of data mimicking the
* source dataset
*
* LexisNexis Risk Solutions patent pending as of June 30, 2020.
*
* Copyright (C) 2024 HPCC Systems
*
* ----------------------------------------------------------------------------
* EXPORTED ATTRIBUTES
* ----------------------------------------------------------------------------
*
* DATA TYPES
* ID_t := UNSIGNED6;
* Hash_t := UNSIGNED8;
* MatchingID_t := UNSIGNED4;
*
* RECORD DEFINITIONS
* LookupTableLayout
* CollapsedMatchingLayout
* IDLayout
* RelatedIDLayout
*
* INDEX FILE DEFINITIONS
* Hash2IDLookupIndexDef(...) := INDEX
* ID2HashLookupIndexDef(...) := INDEX
* Match2IDLookupIndexDef(...) := INDEX
* ID2MatchLookupIndexDef(...) := INDEX
*
* FUNCTIONS (see actual functions for detailed documentation)
* CreateLookupTable(...) := FUNCTIONMACRO
* CollapseLookupTable(...) := FUNCTION
* WriteIDLookupIndexes(...) := FUNCTION
* WriteMatchIDIndexes(...) := FUNCTION
* BuildAllIndexes(...) := FUNCTIONMACRO
* FindRelatedIDs(...) := FUNCTION
* FindRelatedIDsFromLookupTable(...) := FUNCTION
*/
EXPORT Medley := MODULE
//-------------------------------------------------------------------------
// Version information
//-------------------------------------------------------------------------
EXPORT UNSIGNED1 VERSION_MAJOR := 0;
EXPORT UNSIGNED1 VERSION_MINOR := 6;
EXPORT UNSIGNED1 VERSION_POINT := 2;
EXPORT STRING VERSION_STRING := VERSION_MAJOR + '.' + VERSION_MINOR + '.' + VERSION_POINT;
//-------------------------------------------------------------------------
// Data types
//-------------------------------------------------------------------------
EXPORT ID_t := UNSIGNED6;
EXPORT Hash_t := UNSIGNED8;
EXPORT MatchingID_t := UNSIGNED4;
//-------------------------------------------------------------------------
// Exported record definitions
//-------------------------------------------------------------------------
EXPORT LookupTableLayout := RECORD
ID_t id;
Hash_t hash_value;
END;
EXPORT CollapsedMatchingLayout := RECORD
MatchingID_t matching_id;
ID_t id;
END;
EXPORT IDLayout := RECORD
ID_t id;
END;
EXPORT RelatedIDLayout := RECORD
ID_t given_id;
ID_t id;
END;
//-------------------------------------------------------------------------
// Exported index declarations
//-------------------------------------------------------------------------
EXPORT Hash2IDLookupIndexDef(STRING path) := INDEX
(
{LookupTableLayout.hash_value},
{LookupTableLayout},
path
);
EXPORT ID2HashLookupIndexDef(STRING path) := INDEX
(
{LookupTableLayout.id},
{LookupTableLayout},
path
);
EXPORT Match2IDLookupIndexDef(STRING path) := INDEX
(
{CollapsedMatchingLayout.matching_id},
{CollapsedMatchingLayout},
path
);
EXPORT ID2MatchLookupIndexDef(STRING path) := INDEX
(
{CollapsedMatchingLayout.id},
{CollapsedMatchingLayout},
path
);
//-------------------------------------------------------------------------
// Protected record definitions
//-------------------------------------------------------------------------
SHARED MatchIDPairsRec := RECORD
MatchingID_t matching_id;
MatchingID_t related_matching_id;
END;
/**
* Embedded function for reducing linked MatchingID_t pairs. Note that
* reduction occurs on only a per-worker basis; it is possible that
* further reductions are possible in a global process.
*
* The following must be true within the argument:
*
* SORT(SORT(ds, related_matching_id), matching_id, related_matching_id, LOCAL)
* matching_id < related_matching_id for all records
*
* @param ds A DATASET(MatchIDPairsRec) to process
*
* @return A new dataset in the same layout as the argument, with
* all matching_id values reduced as far as possible.
*/
SHARED STREAMED DATASET(MatchIDPairsRec) LocallyReduceMatchPairs(STREAMED DATASET(MatchIDPairsRec) ds) := EMBED(C++ : activity)
#include <map>
#body
typedef uint32_t ID_ELEMENT_TYPE;
typedef std::map<ID_ELEMENT_TYPE, ID_ELEMENT_TYPE> MatchMap;
class MatchIDPairDataset : public RtlCInterface, implements IRowStream
{
public:
MatchIDPairDataset(IEngineRowAllocator* _resultAllocator, IRowStream* _ds)
: resultAllocator(_resultAllocator), matchDataDS(_ds)
{}
RTLIMPLEMENT_IINTERFACE
virtual const void* nextRow() override
{
const byte* oneRow = static_cast<const byte*>(matchDataDS->nextRow());
if (!oneRow)
return nullptr;
ID_ELEMENT_TYPE matchingID = *((ID_ELEMENT_TYPE*)oneRow);
ID_ELEMENT_TYPE relatedMatchingID = *((ID_ELEMENT_TYPE*)(oneRow + sizeof(ID_ELEMENT_TYPE)));
MatchMap::iterator matchingIter = theMap.find(matchingID);
if (matchingIter != theMap.end())
{
matchingID = matchingIter->second;
}
theMap[relatedMatchingID] = matchingID;
RtlDynamicRowBuilder rowBuilder(resultAllocator);
byte* newRow = rowBuilder.getSelf();
*((ID_ELEMENT_TYPE*)newRow) = matchingID;
*((ID_ELEMENT_TYPE*)(newRow + sizeof(ID_ELEMENT_TYPE))) = relatedMatchingID;
// Delete entries we no longer need
for (MatchMap::iterator it = theMap.begin(); it != theMap.end() && it->first < matchingID;)
{
it = theMap.erase(it);
}
return rowBuilder.finalizeRowClear(sizeof(ID_ELEMENT_TYPE) + sizeof(ID_ELEMENT_TYPE));
}
virtual void stop() override
{
matchDataDS->stop();
}
protected:
Linked<IEngineRowAllocator> resultAllocator;
IRowStream* matchDataDS;
MatchMap theMap;
};
return new MatchIDPairDataset(_resultAllocator, ds);
ENDEMBED;
/**
* Function macro that creates deletion neighborhoods for an entire dataset.
* Both inter-field and intra-field deletion neighborhoods may be created,
* depending on the value of the fieldSpec parameter.
*
* @param inFile The dataset to process; REQUIRED
* @param idField The unique identifier field within each
* record; this is not a string; the field's
* data type should match the ID_t definition
* above; REQUIRED
* @param fieldSpec A description of how the dataset should be
* processed; can be a single STRING or a
* SET OF STRING to define several descriptions
* that are combined via OR; within each STRING,
* semi-colons are used to delimit field groups,
* commas are used to delimit fields; field
* names may optionally have a suffix of '%N'
* where N is the maximum edit distance to use
* when creating an inter-field deletion
* neighborhood; fields and field groups may
* have a '&' prefix to indicate that the field
* or field group should not be deleted when
* constructing the intra-field deletion
* neighhborhood; REQUIRED
* @param maxEditDistance The maximum intra-field edit distance to
* create; OPTIONAL, defaults to 1
*
* @return A new DATASET(LookupTableLayout) dataset.
*/
EXPORT CreateLookupTable(inFile, idField, fieldSpec, maxEditDistance = 1) := FUNCTIONMACRO
// Embedded function for creating substrings mapping to a deletion neighborhood;
// note that the strings themselves are returned, not a hash of the string
#UNIQUENAME(CreateStringDeletionNeighborhood);
STREAMED DATASET({UTF8 text}) %CreateStringDeletionNeighborhood%(CONST UTF8 text, UNSIGNED1 max_edit_distance) := EMBED(C++)
#option pure;
#include <set>
#include <string>
#define UCHAR_TYPE uint16_t
#include <unicode/unistr.h>
typedef std::set<std::string> TextSet;
using icu::UnicodeString;
#body
class StreamedStringDataset : public RtlCInterface, implements IRowStream
{
public:
StreamedStringDataset(IEngineRowAllocator* _resultAllocator, unsigned int _word_byte_count, const char* _word, unsigned int _max_edit_distance)
: resultAllocator(_resultAllocator), myText(_word, _word_byte_count, "UTF-8"), myEditDistance(_max_edit_distance), isInited(false)
{
isStopped = false;
}
RTLIMPLEMENT_IINTERFACE
void AppendToCollection(const UnicodeString& textLine)
{
outString.clear();
textLine.toUTF8String(outString);
deletionNeighborhood.insert(outString);
}
void PopulateDeletionNeighborhood(const UnicodeString& textLine, unsigned int depth)
{
if (depth > 0 && textLine.countChar32() > 2)
{
UnicodeString myTextLine;
for (int32_t x = 0; x < textLine.countChar32(); x++)
{
myTextLine = textLine;
myTextLine.remove(x, 1);
AppendToCollection(myTextLine);
PopulateDeletionNeighborhood(myTextLine, depth - 1);
}
}
}
virtual const void* nextRow()
{
if (isStopped)
{
return NULL;
}
if (!isInited)
{
AppendToCollection(myText);
PopulateDeletionNeighborhood(myText, myEditDistance);
deletionNeighborhoodIter = deletionNeighborhood.begin();
isInited = true;
}
if (deletionNeighborhoodIter != deletionNeighborhood.end())
{
const std::string& textLine = *deletionNeighborhoodIter;
RtlDynamicRowBuilder rowBuilder(resultAllocator);
unsigned int len = sizeof(__int32) + textLine.size();
byte* row = rowBuilder.ensureCapacity(len, NULL);
*(__int32*)(row) = rtlUtf8Length(textLine.size(), textLine.data());
memcpy(row + sizeof(__int32), textLine.data(), textLine.size());
++deletionNeighborhoodIter;
return rowBuilder.finalizeRowClear(len);
}
isStopped = true;
return NULL;
}
virtual void stop()
{
isStopped = true;
}
protected:
Linked<IEngineRowAllocator> resultAllocator;
private:
UnicodeString myText;
unsigned int myEditDistance;
TextSet deletionNeighborhood;
TextSet::const_iterator deletionNeighborhoodIter;
std::string outString;
bool isInited;
bool isStopped;
};
return new StreamedStringDataset(_resultAllocator, rtlUtf8Size(lenText, text), text, max_edit_distance);
ENDEMBED;
// Embedded function for creating a numeric deletion neighborhood from a set
// of numbers;
#UNIQUENAME(CreateNumericSetDeletionNeighborhood);
STREAMED DATASET({#$.Medley.Hash_t hash_value}) %CreateNumericSetDeletionNeighborhood%(SET OF #$.Medley.Hash_t _attr_set, UNSIGNED1 max_edit_distance) := EMBED(C++)
#option pure;
#include <set>
#include <vector>
typedef std::vector<hash64_t> AttrList;
typedef std::set<hash64_t> HashSet;
#body
class StreamedHashValueDataset : public RtlCInterface, implements IRowStream
{
public:
StreamedHashValueDataset(IEngineRowAllocator* _resultAllocator, const hash64_t* _set_values, unsigned _num_values, unsigned int _max_edit_distance)
: resultAllocator(_resultAllocator), myEditDistance(_max_edit_distance), isInited(false)
{
isStopped = (_set_values != nullptr && _num_values == 0);
if (!isStopped)
{
for (unsigned int x = 0; x < _num_values; x++)
{
setValues.push_back(_set_values[x]);
}
}
}
RTLIMPLEMENT_IINTERFACE
hash64_t HashStdList(AttrList& myAttrList)
{
hash64_t hashValue = HASH64_INIT;
for (AttrList::const_iterator x = myAttrList.begin(); x != myAttrList.end(); x++)
{
hash64_t element = *x;
hashValue = rtlHash64Data(sizeof(element), &element, hashValue);
}
return hashValue;
}
void PopulateDeletionNeighborhood(AttrList theAttrList, unsigned int depth)
{
if (depth > 0 && theAttrList.size() > 1)
{
AttrList myAttrList(theAttrList.size() - 1, 0); // reserve space
for (unsigned int x = 0; x < theAttrList.size(); x++)
{
if (x > 0)
{
// Copy over the next single element that will not change in
// subsequent loop iterations
myAttrList[x - 1] = theAttrList[x - 1];
}
unsigned int insertPos = x;
// Copy remaining elements
for (unsigned int y = x + 1; y < theAttrList.size(); y++)
{
myAttrList[insertPos++] = theAttrList[y];
}
deletionNeighborhood.insert(HashStdList(myAttrList));
PopulateDeletionNeighborhood(myAttrList, depth - 1);
}
}
}
virtual const void* nextRow()
{
if (isStopped)
{
return NULL;
}
if (!isInited)
{
deletionNeighborhood.insert(HashStdList(setValues));
PopulateDeletionNeighborhood(setValues, myEditDistance);
deletionNeighborhoodIter = deletionNeighborhood.begin();
isInited = true;
}
if (deletionNeighborhoodIter != deletionNeighborhood.end())
{
hash64_t oneHash = *deletionNeighborhoodIter;
RtlDynamicRowBuilder rowBuilder(resultAllocator);
unsigned int len = sizeof(oneHash);
byte* row = rowBuilder.ensureCapacity(len, NULL);
*(hash64_t*)(row) = oneHash;
++deletionNeighborhoodIter;
return rowBuilder.finalizeRowClear(len);
}
isStopped = true;
return NULL;
}
virtual void stop()
{
isStopped = true;
}
protected:
Linked<IEngineRowAllocator> resultAllocator;
private:
AttrList setValues;
unsigned int myEditDistance;
HashSet deletionNeighborhood;
HashSet::const_iterator deletionNeighborhoodIter;
bool isInited;
bool isStopped;
};
hash64_t* setSource = static_cast<hash64_t*>(const_cast<void*>(_attr_set));
unsigned numElements = len_attr_set / sizeof(hash64_t);
return new StreamedHashValueDataset(_resultAllocator, setSource, numElements, (max_edit_distance > numElements - 1 ? numElements - 1 : max_edit_distance));
ENDEMBED;
// Housekeeping involving fieldSpec, which could be a single spec (string)
// or multiple specs (set of strings)
#UNIQUENAME(fieldSpecType);
#SET(fieldSpecType, #GETDATATYPE(fieldSpec));
#UNIQUENAME(fieldSpecIsSet);
#UNIQUENAME(numFieldSpecs);
#IF(%'fieldSpecType'%[..7] = 'set of ')
#SET(fieldSpecIsSet, 1)
#SET(numFieldSpecs, COUNT(fieldSpec))
#ELSE
#SET(fieldSpecIsSet, 0)
#SET(numFieldSpecs, 1)
#END
#UNIQUENAME(fieldSpecIter)
#SET(fieldSpecIter, 1);
// Distribute the incoming data on the ID field, so we can localize operations later
#UNIQUENAME(distInFile);
LOCAL %distInFile% := DISTRIBUTE(inFile, HASH64((#$.Medley.ID_t)idField));
// Make sure edit distance is non-negative
#UNIQUENAME(myMaxEditDistance);
#SET(myMaxEditDistance, (UNSIGNED1)MAX((INTEGER1)maxEditDistance, 0));
// Placeholder for some built-up ECL, combining the results from
// multiple field specs
#UNIQUENAME(combineLookupTableStmt);
#SET(combineLookupTableStmt, '');
#LOOP
#UNIQUENAME(interimLookupTable);
#UNIQUENAME(myFieldSpec)
#UNIQUENAME(requiredFieldSpec)
#UNIQUENAME(requiredFieldSpecCount)
#UNIQUENAME(otherFieldSpec)
#UNIQUENAME(fieldList)
#UNIQUENAME(neighborhoodFields)
#UNIQUENAME(temp)
#UNIQUENAME(tempName)
#UNIQUENAME(tempVal)
#UNIQUENAME(tempSaved)
#UNIQUENAME(pos)
#UNIQUENAME(needDelim)
#IF(%fieldSpecIter% <= %numFieldSpecs%)
// Remove spaces from fieldSpec, insert default edit distance where needed
#IF(%fieldSpecIsSet% = 0)
#SET(myFieldSpec, REGEXREPLACE('%([^\\d])', TRIM((STRING)fieldSpec, ALL), '%1$1'))
#ELSE
#SET(myFieldSpec, REGEXREPLACE('%([^\\d])', TRIM((STRING)(fieldSpec[%fieldSpecIter%]), ALL), '%1$1'))
#END
// Split field spec into required and other
#SET(requiredFieldSpec, '')
#SET(requiredFieldSpecCount, 0)
#SET(otherFieldSpec, '')
#SET(pos, 1)
#LOOP
#SET(temp, REGEXFIND('([^;]+)', %'myFieldSpec'%[%pos%..], 1))
#IF(%'temp'% != '')
#IF(%'temp'%[1] = '&')
#IF(%requiredFieldSpecCount% > 0)
#APPEND(requiredFieldSpec, ';')
#END
#APPEND(requiredFieldSpec, %'temp'%[2..])
#SET(requiredFieldSpecCount, %requiredFieldSpecCount% + 1);
#ELSE
#IF(%'otherFieldSpec'% != '')
#APPEND(otherFieldSpec, ';')
#END
#APPEND(otherFieldSpec, %'temp'%)
#END
#SET(pos, %pos% + LENGTH(%'temp'%) + 1)
#ELSE
#BREAK
#END
#END
// Check for embedded required field group indicators (there should be none)
#IF(REGEXFIND('&', %'requiredFieldSpec'% + %'otherFieldSpec'%))
#ERROR('"' + fieldSpec + '" contains a required indicator (&) on an individual field within a field group')
#END
// Remove any required character patterns from the full field spec
#SET(myFieldSpec, REGEXREPLACE('&', %'myFieldSpec'%, ''))
// Find all fields where we need to create deletion neighborhoods on their values,
// and the maximum edit distance cited (in case the field was defined that way
// more than once)
#SET(neighborhoodFields, ',')
#SET(pos, 1)
#LOOP
#SET(temp, REGEXFIND('([^,;%]+%\\d+)', %'myFieldSpec'%[%pos%..], 1))
#SET(tempName, REGEXFIND('^([^%]+)%', %'temp'%, 1))
#SET(tempVal, REGEXFIND('%(\\d+)', %'temp'%, 1))
#IF(%'tempName'% != '' AND %'tempVal'% != '')
#SET(tempSaved, REGEXFIND('(,' + %'tempName'% + '%\\d+)', %'neighborhoodFields'%, 1))
#IF(%'tempSaved'% != '')
#SET(neighborhoodFields, REGEXREPLACE(',' + %'tempSaved'%, %'neighborhoodFields'%, ',' + %'tempName'% + '%' + (STRING)(MAX((UNSIGNED2)REGEXFIND('%(\\d+)', %'tempSaved'%, 1), (UNSIGNED2)%'tempVal'%))))
#ELSE
#APPEND(neighborhoodFields, %'temp'% + ',')
#END
#SET(pos, %pos% + LENGTH(REGEXFIND('^.+?[^,;%]+%\\d+', %'myFieldSpec'%[%pos%..], 0)) + 1)
#ELSE
#BREAK
#END
#END
#SET(neighborhoodFields, REGEXREPLACE('^,', %'neighborhoodFields'%, ''))
#SET(neighborhoodFields, REGEXREPLACE(',$', %'neighborhoodFields'%, ''))
// Record structure containing only the fields we want, coerced into UTF8 strings
#UNIQUENAME(FileRec)
LOCAL %FileRec% := RECORD
#$.Medley.ID_t id;
#SET(fieldList, '')
#SET(pos, 1)
#LOOP
#SET(temp, REGEXFIND('^([^,;%]+)', %'myFieldSpec'%[%pos%..], 1))
#IF(%'temp'% != '')
#SET(tempSaved, REGEXFIND('^' + %'temp'% + '(%\\d+)?[,;]*', %'myFieldSpec'%[%pos%..], 0))
#IF(NOT REGEXFIND('\\b' + %'temp'% + '\\b', %'fieldList'%))
#IF(%'fieldList'% != '')
#APPEND(fieldList, ',')
#END
#APPEND(fieldList, %'temp'%)
// Include the field in the record definition
UTF8 %temp%;
#END
#SET(pos, %pos% + LENGTH(%'tempSaved'%))
#ELSE
#BREAK
#END
#END
END;
// Create working dataset
#UNIQUENAME(workingFile)
LOCAL %workingFile% := PROJECT
(
%distInFile%,
TRANSFORM
(
%FileRec%,
SELF.id := (#$.Medley.ID_t)LEFT.idField
#SET(pos, 1)
#LOOP
#SET(temp, REGEXFIND('^([^,]+)', %'fieldList'%[%pos%..], 1))
#IF(%'temp'% != '')
, SELF.%temp% := (UTF8)LEFT.%temp%
#SET(pos, %pos% + LENGTH(%'temp'%) + 1)
#ELSE
#BREAK
#END
#END
)
);
// Expand field values with deletion neighborhood entries, if any
#UNIQUENAME(expandedWorkingFile)
#IF(%'neighborhoodFields'% != '')
#UNIQUENAME(resultNameBase)
#UNIQUENAME(resultNameCounter)
#SET(resultNameCounter, 0)
#UNIQUENAME(resultName)
#SET(resultName, %'resultNameBase'% + %'resultNameCounter'%)
#UNIQUENAME(nextResultName)
LOCAL %resultName% := %workingFile%;
#SET(pos, 1)
#LOOP
#SET(temp, REGEXFIND('^([^,]+)', %'neighborhoodFields'%[%pos%..], 1))
#IF(%'temp'% != '')
#SET(tempName, REGEXFIND('^([^%]+)', %'temp'%, 1))
#SET(tempVal, REGEXFIND('%(\\d+)', %'temp'%, 1))
#SET(resultNameCounter, %resultNameCounter% + 1)
#SET(nextResultName, %'resultNameBase'% + %'resultNameCounter'%)
LOCAL %nextResultName% := NORMALIZE
(
%resultName%,
%CreateStringDeletionNeighborhood%(LEFT.%tempName%, %tempVal%),
TRANSFORM
(
RECORDOF(LEFT),
SELF.%tempName% := RIGHT.text,
SELF := LEFT
)
);
#SET(resultName, %'nextResultName'%)
#SET(pos, %pos% + LENGTH(%'temp'%) + 1)
#ELSE
#BREAK
#END
#END
LOCAL %expandedWorkingFile% := %resultName%;
#ELSE
LOCAL %expandedWorkingFile% := %workingFile%;
#END
// Collect all hashes needed for intra-field deletion neighborhood
#UNIQUENAME(requiredHashCmd)
#UNIQUENAME(groupPos)
#UNIQUENAME(groupFields)
#UNIQUENAME(bareFieldNames)
#UNIQUENAME(needsOuterDelim)
#UNIQUENAME(hashSets)
LOCAL %hashSets% := PROJECT
(
%expandedWorkingFile%,
TRANSFORM
(
{
#$.Medley.ID_t id,
#$.Medley.Hash_t required_hash_value,
SET OF #$.Medley.Hash_t hash_values
},
#IF(%requiredFieldSpecCount% > 0)
requiredSet := SET(DATASET
(
[
#SET(groupPos, 1)
#SET(needsOuterDelim, 0)
#SET(bareFieldNames, REGEXREPLACE('%\\d+', %'requiredFieldSpec'%, ''))
#LOOP
#SET(groupFields, REGEXFIND('^([^;]+)', %'bareFieldNames'%[%groupPos%..], 1))
#IF(%'groupFields'% != '')
#IF(%needsOuterDelim% = 1) , #END
HASH64
(
SET(DATASET
(
[
#SET(pos, 1)
#SET(needDelim, 0)
#LOOP
#SET(temp, REGEXFIND('^([^,]+)', %'groupFields'%[%pos%..], 1))
#IF(%'temp'% != '')
#IF(%needDelim% = 1) , #END
IF(TRIM(LEFT.%temp%) != '', %'temp'% + ':' + TRIM(LEFT.%temp%), '')
#SET(needDelim, 1)
#SET(pos, %pos% + LENGTH(%'temp'% + 1))
#ELSE
#BREAK
#END
#END
],
{UTF8 v}
)(v != ''), v)
)
#SET(groupPos, %groupPos% + LENGTH(%'groupFields'% + 1))
#SET(needsOuterDelim, 1)
#ELSE
#BREAK
#END
#END
],
{UNSIGNED8 h}
), h);
#SET(requiredHashCmd, 'HASH64(requiredSet)')
#ELSE
#SET(requiredHashCmd, -1)
#END
otherSet := SET(DATASET
(
[
#SET(groupPos, 1)
#SET(needsOuterDelim, 0)
#SET(bareFieldNames, REGEXREPLACE('%\\d+', %'otherFieldSpec'%, ''))
#LOOP
#SET(groupFields, REGEXFIND('^([^;]+)', %'bareFieldNames'%[%groupPos%..], 1))
#IF(%'groupFields'% != '')
#IF(%needsOuterDelim% = 1) , #END
HASH64
(
SET(DATASET
(
[
#SET(pos, 1)
#SET(needDelim, 0)
#LOOP
#SET(temp, REGEXFIND('^([^,]+)', %'groupFields'%[%pos%..], 1))
#IF(%'temp'% != '')
#IF(%needDelim% = 1) , #END
IF(TRIM(LEFT.%temp%) != U8'', (UTF8)%'temp'% + U8':' + TRIM(LEFT.%temp%), U8'')
#SET(needDelim, 1)
#SET(pos, %pos% + LENGTH(%'temp'% + 1))
#ELSE
#BREAK
#END
#END
],
{UTF8 v}
)(v != U8''), v)
)
#SET(groupPos, %groupPos% + LENGTH(%'groupFields'% + 1))
#SET(needsOuterDelim, 1)
#ELSE
#BREAK
#END
#END
],
{UNSIGNED8 h}
)(h != HASH64([])), h);
SELF.required_hash_value := %requiredHashCmd%,
SELF.hash_values := IF(EXISTS(otherSet), otherSet, [-1]),
SELF := LEFT
)
);
// Apply intra-field deletion neighborhood
#UNIQUENAME(collapsedHashes)
LOCAL %collapsedHashes% := NORMALIZE
(
%hashSets%,
%CreateNumericSetDeletionNeighborhood%(LEFT.hash_values, %myMaxEditDistance%),
TRANSFORM
(
{
#$.Medley.ID_t id,
#$.Medley.Hash_t hash_value
},
SELF.hash_value := HASH64(LEFT.required_hash_value, RIGHT.hash_value),
SELF := LEFT
)
);
// Assign deduped data to an interim attribute
LOCAL %interimLookupTable% := DEDUP(SORT(%collapsedHashes%, id, hash_value, LOCAL), id, hash_value, LOCAL)
// Append the iterim attribute to our collection of attributes
#IF(%fieldSpecIter% > 1)
#APPEND(combineLookupTableStmt, '+')
#END
#APPEND(combineLookupTableStmt, %'interimLookupTable'%)
#SET(fieldSpecIter, %fieldSpecIter% + 1)
#ELSE
#BREAK
#END
#END
#UNIQUENAME(finalResult);
LOCAL %finalResult% :=
#IF(%fieldSpecIsSet% = 0)
%combineLookupTableStmt%
#ELSE
// Dedup the collected interim attributes
DEDUP(SORT(%combineLookupTableStmt%, id, hash_value, LOCAL), id, hash_value, LOCAL)
#END;
RETURN %finalResult%;
ENDMACRO;
/**
* Create ID <-> hash lookup index files from a lookup table (as returned
* by CreateLookupTable()). This function will overwrite existing
* files with the same path.
*
* @param lookupTable The lookup table (as returned by
* CreateLookupTable()); REQUIRED
* @param id2HashIndexPath Logical pathname of the ID->Hash
* index file; REQUIRED
* @param hash2IDIndexPath Logical pathname of the Hash->ID
* index file; REQUIRED
*
* @return An action that creates the two index files.
*
* @see CreateLookupTable
*/
EXPORT WriteIDLookupIndexes(DATASET(LookupTableLayout) lookupTable,
STRING id2HashIndexPath,
STRING hash2IDIndexPath) := FUNCTION
RETURN PARALLEL
(
BUILD(Hash2IDLookupIndexDef(hash2IDIndexPath), lookupTable, OVERWRITE);
BUILD(ID2HashLookupIndexDef(id2HashIndexPath), lookupTable, OVERWRITE);
);
END;
/**
* Performs a "fuzzy deduplication" from a lookup table (as if returned from
* CreateLookupTable()).
*
* @param lookupTable Dataset to process
*
* @return A new DATASET(CollapsedMatchingLayout).
*
* @see CreateLookupTable
*/
EXPORT DATASET(CollapsedMatchingLayout) CollapseLookupTable(DATASET(LookupTableLayout) lookupTable) := FUNCTION
// Prep for rollup
idLinksAsSet0 := PROJECT
(
lookupTable,
TRANSFORM
(
{
RECORDOF(LEFT),
SET OF ID_t id_set
},
SELF.id_set := [LEFT.id],
SELF := LEFT
)
);
// Group all shared IDs under the same hash value
idLinksAsSet := ROLLUP
(
SORT(idLinksAsSet0, hash_value, id),
TRANSFORM
(
RECORDOF(LEFT),
SELF.id_set := LEFT.id_set + RIGHT.id_set,
SELF := LEFT
),
hash_value
);
// ... and assign an initial matching_id to them
matchedIDSets := PROJECT
(
idLinksAsSet,
TRANSFORM
(
{
MatchingID_t matching_id,
SET OF ID_t id_set
},
SELF.matching_id := COUNTER,
SELF := LEFT
)
);
// Break the sets back out again, which basically gives us
// a tiny matching_id/entity_id pair
normalizedIDMatchID := NORMALIZE
(
matchedIDSets,
DATASET(LEFT.id_set, {ID_t id}),
TRANSFORM
(
{
MatchingID_t matching_id,
ID_t id,
MatchingID_t related_matching_id
},
SELF.matching_id := LEFT.matching_id,
SELF.id := RIGHT.id,
SELF.related_matching_id := 0 // will assign later
)
);
// Initial 'glue' of rewriting one matching_id to a lower matching_id
// on a per-entity ID basis; this forms a chain from a matching_id
// to lower-valued IDs
idSetsWithRelated := ITERATE
(
SORT(normalizedIDMatchID, id, -matching_id),
TRANSFORM
(
RECORDOF(LEFT),
SELF.related_matching_id := IF(LEFT.id = RIGHT.id, LEFT.matching_id, 0),
SELF := RIGHT
)
);
// Further reduction: Make sure each related_matching_id references
// the lower values, in case of duplicates
uniqueMatchIDPairs1 := TABLE
(
idSetsWithRelated(related_matching_id != 0),
{
related_matching_id,
MatchingID_t matching_id := MIN(GROUP, matching_id)
},
related_matching_id,
MERGE
);
// Prepare this dataset for per-worker "chain walking" to minimize
// matching_id for each related_matching_id; if the dataset is small
// enough, distribute it to a single node so the subsequent LOOP has
// less work to do
uniqueMatchIDPairs2 := uniqueMatchIDPairs1(matching_id < related_matching_id);
uniqueMatchIDPairsLarge := SORT(uniqueMatchIDPairs2, related_matching_id);
uniqueMatchIDPairsSmall := DISTRIBUTE(uniqueMatchIDPairs2, 1);
uniqueMatchIDPairs3 := IF(COUNT(uniqueMatchIDPairs2) < 1000000, uniqueMatchIDPairsSmall, uniqueMatchIDPairsLarge);
uniqueMatchIDPairs4 := SORT(uniqueMatchIDPairs3, matching_id, related_matching_id, LOCAL);
uniqueMatchIDPairs5 := PROJECT(uniqueMatchIDPairs4, MatchIDPairsRec);
uniqueMatchIDPairs := uniqueMatchIDPairs5;
reducedUniqueMatchIDPairs := LocallyReduceMatchPairs(uniqueMatchIDPairs);
// Total reduction is not possible with a large reducedUniqueMatchIDPairs
// because LocallyReduceMatchPairs() works only on a worker's local data
// and chains could span workers; prepare the data for LOOP
tempNormalizedIDMatchID := PROJECT
(
normalizedIDMatchID,
TRANSFORM
(
{
RECORDOF(LEFT),
BOOLEAN needsUpdating
},
SELF.needsUpdating := TRUE,
SELF := LEFT
)
);
// Repeatedly rewrite the original matching_id value, stopping only
// when we no longer have chain links to walk
reducedIDMatchID := LOOP