1/* Copyright (c) 2001, 2010, Oracle and/or its affiliates.
2 Copyright (c) 2010, 2015, MariaDB
3
4 This program is free software; you can redistribute it and/or modify
5 it under the terms of the GNU General Public License as published by
6 the Free Software Foundation; version 2 of the License.
7
8 This program is distributed in the hope that it will be useful,
9 but WITHOUT ANY WARRANTY; without even the implied warranty of
10 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
11 GNU General Public License for more details.
12
13 You should have received a copy of the GNU General Public License
14 along with this program; if not, write to the Free Software
15 Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA */
16
17/*
18 Function to handle quick removal of duplicates
19 This code is used when doing multi-table deletes to find the rows in
20 reference tables that needs to be deleted.
21
22 The basic idea is as follows:
23
24 Store first all strings in a binary tree, ignoring duplicates.
25 When the tree uses more memory than 'max_heap_table_size',
26 write the tree (in sorted order) out to disk and start with a new tree.
27 When all data has been generated, merge the trees (removing any found
28 duplicates).
29
30 The unique entries will be returned in sort order, to ensure that we do the
31 deletes in disk order.
32*/
33
34#include "mariadb.h"
35#include "sql_priv.h"
36#include "unireg.h"
37#include "sql_sort.h"
38#include "queues.h" // QUEUE
39#include "my_tree.h" // element_count
40#include "uniques.h" // Unique
41#include "sql_sort.h"
42#include "myisamchk.h" // BUFFPEK
43
44int unique_write_to_file(uchar* key, element_count count, Unique *unique)
45{
46 /*
47 Use unique->size (size of element stored in the tree) and not
48 unique->tree.size_of_element. The latter is different from unique->size
49 when tree implementation chooses to store pointer to key in TREE_ELEMENT
50 (instead of storing the element itself there)
51 */
52 return my_b_write(&unique->file, key, unique->size) ? 1 : 0;
53}
54
55int unique_write_to_file_with_count(uchar* key, element_count count, Unique *unique)
56{
57 return my_b_write(&unique->file, key, unique->size) ||
58 my_b_write(&unique->file, (uchar*)&count, sizeof(element_count)) ? 1 : 0;
59}
60
61int unique_write_to_ptrs(uchar* key, element_count count, Unique *unique)
62{
63 memcpy(unique->sort.record_pointers, key, unique->size);
64 unique->sort.record_pointers+=unique->size;
65 return 0;
66}
67
68int unique_intersect_write_to_ptrs(uchar* key, element_count count, Unique *unique)
69{
70 if (count >= unique->min_dupl_count)
71 {
72 memcpy(unique->sort.record_pointers, key, unique->size);
73 unique->sort.record_pointers+=unique->size;
74 }
75 else
76 unique->filtered_out_elems++;
77 return 0;
78}
79
80
81Unique::Unique(qsort_cmp2 comp_func, void * comp_func_fixed_arg,
82 uint size_arg, size_t max_in_memory_size_arg,
83 uint min_dupl_count_arg)
84 :max_in_memory_size(max_in_memory_size_arg),
85 size(size_arg),
86 elements(0)
87{
88 my_b_clear(&file);
89 min_dupl_count= min_dupl_count_arg;
90 full_size= size;
91 if (min_dupl_count_arg)
92 full_size+= sizeof(element_count);
93 with_counters= MY_TEST(min_dupl_count_arg);
94 init_tree(&tree, (max_in_memory_size / 16), 0, size, comp_func,
95 NULL, comp_func_fixed_arg, MYF(MY_THREAD_SPECIFIC));
96 /* If the following fail's the next add will also fail */
97 my_init_dynamic_array(&file_ptrs, sizeof(BUFFPEK), 16, 16,
98 MYF(MY_THREAD_SPECIFIC));
99 /*
100 If you change the following, change it in get_max_elements function, too.
101 */
102 max_elements= (ulong) (max_in_memory_size /
103 ALIGN_SIZE(sizeof(TREE_ELEMENT)+size));
104 if (!max_elements)
105 max_elements= 1;
106
107 (void) open_cached_file(&file, mysql_tmpdir,TEMP_PREFIX, DISK_BUFFER_SIZE,
108 MYF(MY_WME));
109}
110
111
112/*
113 Calculate log2(n!)
114
115 NOTES
116 Stirling's approximate formula is used:
117
118 n! ~= sqrt(2*M_PI*n) * (n/M_E)^n
119
120 Derivation of formula used for calculations is as follows:
121
122 log2(n!) = log(n!)/log(2) = log(sqrt(2*M_PI*n)*(n/M_E)^n) / log(2) =
123
124 = (log(2*M_PI*n)/2 + n*log(n/M_E)) / log(2).
125*/
126
127inline double log2_n_fact(double x)
128{
129 return (log(2*M_PI*x)/2 + x*log(x/M_E)) / M_LN2;
130}
131
132
133/*
134 Calculate cost of merge_buffers function call for given sequence of
135 input stream lengths and store the number of rows in result stream in *last.
136
137 SYNOPSIS
138 get_merge_buffers_cost()
139 buff_elems Array of #s of elements in buffers
140 elem_size Size of element stored in buffer
141 first Pointer to first merged element size
142 last Pointer to last merged element size
143
144 RETURN
145 Cost of merge_buffers operation in disk seeks.
146
147 NOTES
148 It is assumed that no rows are eliminated during merge.
149 The cost is calculated as
150
151 cost(read_and_write) + cost(merge_comparisons).
152
153 All bytes in the sequences is read and written back during merge so cost
154 of disk io is 2*elem_size*total_buf_elems/IO_SIZE (2 is for read + write)
155
156 For comparisons cost calculations we assume that all merged sequences have
157 the same length, so each of total_buf_size elements will be added to a sort
158 heap with (n_buffers-1) elements. This gives the comparison cost:
159
160 total_buf_elems* log2(n_buffers) / TIME_FOR_COMPARE_ROWID;
161*/
162
163static double get_merge_buffers_cost(uint *buff_elems, uint elem_size,
164 uint *first, uint *last,
165 uint compare_factor)
166{
167 uint total_buf_elems= 0;
168 for (uint *pbuf= first; pbuf <= last; pbuf++)
169 total_buf_elems+= *pbuf;
170 *last= total_buf_elems;
171
172 size_t n_buffers= last - first + 1;
173
174 /* Using log2(n)=log(n)/log(2) formula */
175 return 2*((double)total_buf_elems*elem_size) / IO_SIZE +
176 total_buf_elems*log((double) n_buffers) / (compare_factor * M_LN2);
177}
178
179
180/*
181 Calculate cost of merging buffers into one in Unique::get, i.e. calculate
182 how long (in terms of disk seeks) the two calls
183 merge_many_buffs(...);
184 merge_buffers(...);
185 will take.
186
187 SYNOPSIS
188 get_merge_many_buffs_cost()
189 buffer buffer space for temporary data, at least
190 Unique::get_cost_calc_buff_size bytes
191 maxbuffer # of full buffers
192 max_n_elems # of elements in first maxbuffer buffers
193 last_n_elems # of elements in last buffer
194 elem_size size of buffer element
195
196 NOTES
197 maxbuffer+1 buffers are merged, where first maxbuffer buffers contain
198 max_n_elems elements each and last buffer contains last_n_elems elements.
199
200 The current implementation does a dumb simulation of merge_many_buffs
201 function actions.
202
203 RETURN
204 Cost of merge in disk seeks.
205*/
206
207static double get_merge_many_buffs_cost(uint *buffer,
208 uint maxbuffer, uint max_n_elems,
209 uint last_n_elems, int elem_size,
210 uint compare_factor)
211{
212 int i;
213 double total_cost= 0.0;
214 uint *buff_elems= buffer; /* #s of elements in each of merged sequences */
215
216 /*
217 Set initial state: first maxbuffer sequences contain max_n_elems elements
218 each, last sequence contains last_n_elems elements.
219 */
220 for (i = 0; i < (int)maxbuffer; i++)
221 buff_elems[i]= max_n_elems;
222 buff_elems[maxbuffer]= last_n_elems;
223
224 /*
225 Do it exactly as merge_many_buff function does, calling
226 get_merge_buffers_cost to get cost of merge_buffers.
227 */
228 if (maxbuffer >= MERGEBUFF2)
229 {
230 while (maxbuffer >= MERGEBUFF2)
231 {
232 uint lastbuff= 0;
233 for (i = 0; i <= (int) maxbuffer - MERGEBUFF*3/2; i += MERGEBUFF)
234 {
235 total_cost+=get_merge_buffers_cost(buff_elems, elem_size,
236 buff_elems + i,
237 buff_elems + i + MERGEBUFF-1,
238 compare_factor);
239 lastbuff++;
240 }
241 total_cost+=get_merge_buffers_cost(buff_elems, elem_size,
242 buff_elems + i,
243 buff_elems + maxbuffer,
244 compare_factor);
245 maxbuffer= lastbuff;
246 }
247 }
248
249 /* Simulate final merge_buff call. */
250 total_cost += get_merge_buffers_cost(buff_elems, elem_size,
251 buff_elems, buff_elems + maxbuffer,
252 compare_factor);
253 return total_cost;
254}
255
256
257/*
258 Calculate cost of using Unique for processing nkeys elements of size
259 key_size using max_in_memory_size memory.
260
261 SYNOPSIS
262 Unique::get_use_cost()
263 buffer space for temporary data, use Unique::get_cost_calc_buff_size
264 to get # bytes needed.
265 nkeys #of elements in Unique
266 key_size size of each elements in bytes
267 max_in_memory_size amount of memory Unique will be allowed to use
268 compare_factor used to calculate cost of one comparison
269 write_fl if the result must be saved written to disk
270 in_memory_elems OUT estimate of the number of elements in memory
271 if disk is not used
272
273 RETURN
274 Cost in disk seeks.
275
276 NOTES
277 cost(using_unqiue) =
278 cost(create_trees) + (see #1)
279 cost(merge) + (see #2)
280 cost(read_result) (see #3)
281
282 1. Cost of trees creation
283 For each Unique::put operation there will be 2*log2(n+1) elements
284 comparisons, where n runs from 1 tree_size (we assume that all added
285 elements are different). Together this gives:
286
287 n_compares = 2*(log2(2) + log2(3) + ... + log2(N+1)) = 2*log2((N+1)!)
288
289 then cost(tree_creation) = n_compares*ROWID_COMPARE_COST;
290
291 Total cost of creating trees:
292 (n_trees - 1)*max_size_tree_cost + non_max_size_tree_cost.
293
294 Approximate value of log2(N!) is calculated by log2_n_fact function.
295
296 2. Cost of merging.
297 If only one tree is created by Unique no merging will be necessary.
298 Otherwise, we model execution of merge_many_buff function and count
299 #of merges. (The reason behind this is that number of buffers is small,
300 while size of buffers is big and we don't want to loose precision with
301 O(x)-style formula)
302
303 3. If only one tree is created by Unique no disk io will happen.
304 Otherwise, ceil(key_len*n_keys) disk seeks are necessary. We assume
305 these will be random seeks.
306*/
307
308double Unique::get_use_cost(uint *buffer, size_t nkeys, uint key_size,
309 size_t max_in_memory_size,
310 uint compare_factor,
311 bool intersect_fl, bool *in_memory)
312{
313 size_t max_elements_in_tree;
314 size_t last_tree_elems;
315 size_t n_full_trees; /* number of trees in unique - 1 */
316 double result;
317
318 max_elements_in_tree= ((size_t) max_in_memory_size /
319 ALIGN_SIZE(sizeof(TREE_ELEMENT)+key_size));
320
321 n_full_trees= nkeys / max_elements_in_tree;
322 last_tree_elems= nkeys % max_elements_in_tree;
323
324 /* Calculate cost of creating trees */
325 result= 2*log2_n_fact(last_tree_elems + 1.0);
326 if (n_full_trees)
327 result+= n_full_trees * log2_n_fact(max_elements_in_tree + 1.0);
328 result /= compare_factor;
329
330 DBUG_PRINT("info",("unique trees sizes: %u=%u*%u + %u", (uint)nkeys,
331 (uint)n_full_trees,
332 (uint)(n_full_trees?max_elements_in_tree:0),
333 (uint)last_tree_elems));
334
335 if (in_memory)
336 *in_memory= !n_full_trees;
337
338 if (!n_full_trees)
339 return result;
340
341 /*
342 There is more then one tree and merging is necessary.
343 First, add cost of writing all trees to disk, assuming that all disk
344 writes are sequential.
345 */
346 result += DISK_SEEK_BASE_COST * n_full_trees *
347 ceil(((double) key_size)*max_elements_in_tree / IO_SIZE);
348 result += DISK_SEEK_BASE_COST * ceil(((double) key_size)*last_tree_elems / IO_SIZE);
349
350 /* Cost of merge */
351 if (intersect_fl)
352 key_size+= sizeof(element_count);
353 double merge_cost= get_merge_many_buffs_cost(buffer, (uint)n_full_trees,
354 (uint)max_elements_in_tree,
355 (uint)last_tree_elems, key_size,
356 compare_factor);
357 result += merge_cost;
358 /*
359 Add cost of reading the resulting sequence, assuming there were no
360 duplicate elements.
361 */
362 result += ceil((double)key_size*nkeys/IO_SIZE);
363
364 return result;
365}
366
367Unique::~Unique()
368{
369 close_cached_file(&file);
370 delete_tree(&tree, 0);
371 delete_dynamic(&file_ptrs);
372}
373
374
375 /* Write tree to disk; clear tree */
376bool Unique::flush()
377{
378 BUFFPEK file_ptr;
379 elements+= tree.elements_in_tree;
380 file_ptr.count=tree.elements_in_tree;
381 file_ptr.file_pos=my_b_tell(&file);
382
383 tree_walk_action action= min_dupl_count ?
384 (tree_walk_action) unique_write_to_file_with_count :
385 (tree_walk_action) unique_write_to_file;
386 if (tree_walk(&tree, action,
387 (void*) this, left_root_right) ||
388 insert_dynamic(&file_ptrs, (uchar*) &file_ptr))
389 return 1;
390 delete_tree(&tree, 0);
391 return 0;
392}
393
394
395/*
396 Clear the tree and the file.
397 You must call reset() if you want to reuse Unique after walk().
398*/
399
400void
401Unique::reset()
402{
403 reset_tree(&tree);
404 /*
405 If elements != 0, some trees were stored in the file (see how
406 flush() works). Note, that we can not count on my_b_tell(&file) == 0
407 here, because it can return 0 right after walk(), and walk() does not
408 reset any Unique member.
409 */
410 if (elements)
411 {
412 reset_dynamic(&file_ptrs);
413 reinit_io_cache(&file, WRITE_CACHE, 0L, 0, 1);
414 }
415 my_free(sort.record_pointers);
416 elements= 0;
417 tree.flag= 0;
418 sort.record_pointers= 0;
419}
420
421/*
422 The comparison function, passed to queue_init() in merge_walk() and in
423 merge_buffers() when the latter is called from Uniques::get() must
424 use comparison function of Uniques::tree, but compare members of struct
425 BUFFPEK.
426*/
427
428C_MODE_START
429
430static int buffpek_compare(void *arg, uchar *key_ptr1, uchar *key_ptr2)
431{
432 BUFFPEK_COMPARE_CONTEXT *ctx= (BUFFPEK_COMPARE_CONTEXT *) arg;
433 return ctx->key_compare(ctx->key_compare_arg,
434 *((uchar **) key_ptr1), *((uchar **)key_ptr2));
435}
436
437C_MODE_END
438
439
440inline
441element_count get_counter_from_merged_element(void *ptr, uint ofs)
442{
443 element_count cnt;
444 memcpy((uchar *) &cnt, (uchar *) ptr + ofs, sizeof(element_count));
445 return cnt;
446}
447
448
449inline
450void put_counter_into_merged_element(void *ptr, uint ofs, element_count cnt)
451{
452 memcpy((uchar *) ptr + ofs, (uchar *) &cnt, sizeof(element_count));
453}
454
455
456/*
457 DESCRIPTION
458
459 Function is very similar to merge_buffers, but instead of writing sorted
460 unique keys to the output file, it invokes walk_action for each key.
461 This saves I/O if you need to pass through all unique keys only once.
462
463 SYNOPSIS
464 merge_walk()
465 All params are 'IN' (but see comment for begin, end):
466 merge_buffer buffer to perform cached piece-by-piece loading
467 of trees; initially the buffer is empty
468 merge_buffer_size size of merge_buffer. Must be aligned with
469 key_length
470 key_length size of tree element; key_length * (end - begin)
471 must be less or equal than merge_buffer_size.
472 begin pointer to BUFFPEK struct for the first tree.
473 end pointer to BUFFPEK struct for the last tree;
474 end > begin and [begin, end) form a consecutive
475 range. BUFFPEKs structs in that range are used and
476 overwritten in merge_walk().
477 walk_action element visitor. Action is called for each unique
478 key.
479 walk_action_arg argument to walk action. Passed to it on each call.
480 compare elements comparison function
481 compare_arg comparison function argument
482 file file with all trees dumped. Trees in the file
483 must contain sorted unique values. Cache must be
484 initialized in read mode.
485 with counters take into account counters for equal merged
486 elements
487 RETURN VALUE
488 0 ok
489 <> 0 error
490*/
491
492static bool merge_walk(uchar *merge_buffer, size_t merge_buffer_size,
493 uint key_length, BUFFPEK *begin, BUFFPEK *end,
494 tree_walk_action walk_action, void *walk_action_arg,
495 qsort_cmp2 compare, void *compare_arg,
496 IO_CACHE *file, bool with_counters)
497{
498 BUFFPEK_COMPARE_CONTEXT compare_context = { compare, compare_arg };
499 QUEUE queue;
500 if (end <= begin ||
501 merge_buffer_size < (size_t) (key_length * (end - begin + 1)) ||
502 init_queue(&queue, (uint) (end - begin), offsetof(BUFFPEK, key), 0,
503 buffpek_compare, &compare_context, 0, 0))
504 return 1;
505 /* we need space for one key when a piece of merge buffer is re-read */
506 merge_buffer_size-= key_length;
507 uchar *save_key_buff= merge_buffer + merge_buffer_size;
508 uint max_key_count_per_piece= (uint) (merge_buffer_size/(end-begin) /
509 key_length);
510 /* if piece_size is aligned reuse_freed_buffer will always hit */
511 uint piece_size= max_key_count_per_piece * key_length;
512 ulong bytes_read; /* to hold return value of read_to_buffer */
513 BUFFPEK *top;
514 int res= 1;
515 uint cnt_ofs= key_length - (with_counters ? sizeof(element_count) : 0);
516 element_count cnt;
517 /*
518 Invariant: queue must contain top element from each tree, until a tree
519 is not completely walked through.
520 Here we're forcing the invariant, inserting one element from each tree
521 to the queue.
522 */
523 for (top= begin; top != end; ++top)
524 {
525 top->base= merge_buffer + (top - begin) * piece_size;
526 top->max_keys= max_key_count_per_piece;
527 bytes_read= read_to_buffer(file, top, key_length);
528 if (unlikely(bytes_read == (ulong) -1))
529 goto end;
530 DBUG_ASSERT(bytes_read);
531 queue_insert(&queue, (uchar *) top);
532 }
533 top= (BUFFPEK *) queue_top(&queue);
534 while (queue.elements > 1)
535 {
536 /*
537 Every iteration one element is removed from the queue, and one is
538 inserted by the rules of the invariant. If two adjacent elements on
539 the top of the queue are not equal, biggest one is unique, because all
540 elements in each tree are unique. Action is applied only to unique
541 elements.
542 */
543 void *old_key= top->key;
544 /*
545 read next key from the cache or from the file and push it to the
546 queue; this gives new top.
547 */
548 top->key+= key_length;
549 if (--top->mem_count)
550 queue_replace_top(&queue);
551 else /* next piece should be read */
552 {
553 /* save old_key not to overwrite it in read_to_buffer */
554 memcpy(save_key_buff, old_key, key_length);
555 old_key= save_key_buff;
556 bytes_read= read_to_buffer(file, top, key_length);
557 if (unlikely(bytes_read == (ulong) -1))
558 goto end;
559 else if (bytes_read) /* top->key, top->mem_count are reset */
560 queue_replace_top(&queue); /* in read_to_buffer */
561 else
562 {
563 /*
564 Tree for old 'top' element is empty: remove it from the queue and
565 give all its memory to the nearest tree.
566 */
567 queue_remove_top(&queue);
568 reuse_freed_buff(&queue, top, key_length);
569 }
570 }
571 top= (BUFFPEK *) queue_top(&queue);
572 /* new top has been obtained; if old top is unique, apply the action */
573 if (compare(compare_arg, old_key, top->key))
574 {
575 cnt= with_counters ?
576 get_counter_from_merged_element(old_key, cnt_ofs) : 1;
577 if (walk_action(old_key, cnt, walk_action_arg))
578 goto end;
579 }
580 else if (with_counters)
581 {
582 cnt= get_counter_from_merged_element(top->key, cnt_ofs);
583 cnt+= get_counter_from_merged_element(old_key, cnt_ofs);
584 put_counter_into_merged_element(top->key, cnt_ofs, cnt);
585 }
586 }
587 /*
588 Applying walk_action to the tail of the last tree: this is safe because
589 either we had only one tree in the beginning, either we work with the
590 last tree in the queue.
591 */
592 do
593 {
594 do
595 {
596
597 cnt= with_counters ?
598 get_counter_from_merged_element(top->key, cnt_ofs) : 1;
599 if (walk_action(top->key, cnt, walk_action_arg))
600 goto end;
601 top->key+= key_length;
602 }
603 while (--top->mem_count);
604 bytes_read= read_to_buffer(file, top, key_length);
605 if (unlikely(bytes_read == (ulong) -1))
606 goto end;
607 }
608 while (bytes_read);
609 res= 0;
610end:
611 delete_queue(&queue);
612 return res;
613}
614
615
616/*
617 DESCRIPTION
618 Walks consecutively through all unique elements:
619 if all elements are in memory, then it simply invokes 'tree_walk', else
620 all flushed trees are loaded to memory piece-by-piece, pieces are
621 sorted, and action is called for each unique value.
622 Note: so as merging resets file_ptrs state, this method can change
623 internal Unique state to undefined: if you want to reuse Unique after
624 walk() you must call reset() first!
625 SYNOPSIS
626 Unique:walk()
627 All params are 'IN':
628 table parameter for the call of the merge method
629 action function-visitor, typed in include/my_tree.h
630 function is called for each unique element
631 arg argument for visitor, which is passed to it on each call
632 RETURN VALUE
633 0 OK
634 <> 0 error
635 */
636
637bool Unique::walk(TABLE *table, tree_walk_action action, void *walk_action_arg)
638{
639 int res= 0;
640 uchar *merge_buffer;
641
642 if (elements == 0) /* the whole tree is in memory */
643 return tree_walk(&tree, action, walk_action_arg, left_root_right);
644
645 sort.return_rows= elements+tree.elements_in_tree;
646 /* flush current tree to the file to have some memory for merge buffer */
647 if (flush())
648 return 1;
649 if (flush_io_cache(&file) || reinit_io_cache(&file, READ_CACHE, 0L, 0, 0))
650 return 1;
651 /*
652 merge_buffer must fit at least MERGEBUFF2 + 1 keys, because
653 merge_index() can merge that many BUFFPEKs at once. The extra space for one key
654 is needed when a piece of merge buffer is re-read, see merge_walk()
655 */
656 size_t buff_sz= MY_MAX(MERGEBUFF2+1, max_in_memory_size/full_size+1) * full_size;
657 if (!(merge_buffer = (uchar *)my_malloc(buff_sz, MYF(MY_WME))))
658 return 1;
659 if (buff_sz < full_size * (file_ptrs.elements + 1UL))
660 res= merge(table, merge_buffer, buff_sz >= full_size * MERGEBUFF2) ;
661
662 if (!res)
663 {
664 res= merge_walk(merge_buffer, buff_sz, full_size,
665 (BUFFPEK *) file_ptrs.buffer,
666 (BUFFPEK *) file_ptrs.buffer + file_ptrs.elements,
667 action, walk_action_arg,
668 tree.compare, tree.custom_arg, &file, with_counters);
669 }
670 my_free(merge_buffer);
671 return res;
672}
673
674
675/*
676 DESCRIPTION
677
678 Perform multi-pass sort merge of the elements using the buffer buff as
679 the merge buffer. The last pass is not performed if without_last_merge is
680 TRUE.
681
682 SYNOPSIS
683 Unique:merge()
684 All params are 'IN':
685 table the parameter to access sort context
686 buff merge buffer
687 without_last_merge TRUE <=> do not perform the last merge
688 RETURN VALUE
689 0 OK
690 <> 0 error
691 */
692
693bool Unique::merge(TABLE *table, uchar *buff, bool without_last_merge)
694{
695 IO_CACHE *outfile= &sort.io_cache;
696 BUFFPEK *file_ptr= (BUFFPEK*) file_ptrs.buffer;
697 uint maxbuffer= file_ptrs.elements - 1;
698 my_off_t save_pos;
699 bool error= 1;
700 Sort_param sort_param;
701
702 /* Open cached file for table records if it isn't open */
703 if (! my_b_inited(outfile) &&
704 open_cached_file(outfile,mysql_tmpdir,TEMP_PREFIX,READ_RECORD_BUFFER,
705 MYF(MY_WME)))
706 return 1;
707
708 bzero((char*) &sort_param,sizeof(sort_param));
709 sort_param.max_rows= elements;
710 sort_param.sort_form= table;
711 sort_param.rec_length= sort_param.sort_length= sort_param.ref_length=
712 full_size;
713 sort_param.min_dupl_count= min_dupl_count;
714 sort_param.res_length= 0;
715 sort_param.max_keys_per_buffer=
716 (uint) MY_MAX((max_in_memory_size / sort_param.sort_length), MERGEBUFF2);
717 sort_param.not_killable= 1;
718
719 sort_param.unique_buff= buff +(sort_param.max_keys_per_buffer *
720 sort_param.sort_length);
721
722 sort_param.compare= (qsort2_cmp) buffpek_compare;
723 sort_param.cmp_context.key_compare= tree.compare;
724 sort_param.cmp_context.key_compare_arg= tree.custom_arg;
725
726 /* Merge the buffers to one file, removing duplicates */
727 if (merge_many_buff(&sort_param,buff,file_ptr,&maxbuffer,&file))
728 goto err;
729 if (flush_io_cache(&file) ||
730 reinit_io_cache(&file,READ_CACHE,0L,0,0))
731 goto err;
732 sort_param.res_length= sort_param.rec_length-
733 (min_dupl_count ? sizeof(min_dupl_count) : 0);
734 if (without_last_merge)
735 {
736 file_ptrs.elements= maxbuffer+1;
737 return 0;
738 }
739 if (merge_index(&sort_param, buff, file_ptr, maxbuffer, &file, outfile))
740 goto err;
741 error= 0;
742err:
743 if (flush_io_cache(outfile))
744 error= 1;
745
746 /* Setup io_cache for reading */
747 save_pos= outfile->pos_in_file;
748 if (reinit_io_cache(outfile,READ_CACHE,0L,0,0))
749 error= 1;
750 outfile->end_of_file=save_pos;
751 return error;
752}
753
754
755/*
756 Allocate memory that can be used with init_records() so that
757 rows will be read in priority order.
758*/
759
760bool Unique::get(TABLE *table)
761{
762 bool rc= 1;
763 uchar *sort_buffer= NULL;
764 sort.return_rows= elements+tree.elements_in_tree;
765 DBUG_ENTER("Unique::get");
766
767 if (my_b_tell(&file) == 0)
768 {
769 /* Whole tree is in memory; Don't use disk if you don't need to */
770 if ((sort.record_pointers= (uchar*)
771 my_malloc(size * tree.elements_in_tree, MYF(MY_THREAD_SPECIFIC))))
772 {
773 uchar *save_record_pointers= sort.record_pointers;
774 tree_walk_action action= min_dupl_count ?
775 (tree_walk_action) unique_intersect_write_to_ptrs :
776 (tree_walk_action) unique_write_to_ptrs;
777 filtered_out_elems= 0;
778 (void) tree_walk(&tree, action,
779 this, left_root_right);
780 /* Restore record_pointers that was changed in by 'action' above */
781 sort.record_pointers= save_record_pointers;
782 sort.return_rows-= filtered_out_elems;
783 DBUG_RETURN(0);
784 }
785 }
786 /* Not enough memory; Save the result to file && free memory used by tree */
787 if (flush())
788 DBUG_RETURN(1);
789 size_t buff_sz= (max_in_memory_size / full_size + 1) * full_size;
790 if (!(sort_buffer= (uchar*) my_malloc(buff_sz,
791 MYF(MY_THREAD_SPECIFIC|MY_WME))))
792 DBUG_RETURN(1);
793
794 if (merge(table, sort_buffer, FALSE))
795 goto err;
796 rc= 0;
797
798err:
799 my_free(sort_buffer);
800 DBUG_RETURN(rc);
801}
802