Line data Source code
1 : /*-------------------------------------------------------------------------
2 : *
3 : * bernoulli.c
4 : * support routines for BERNOULLI tablesample method
5 : *
6 : * To ensure repeatability of samples, it is necessary that selection of a
7 : * given tuple be history-independent; otherwise syncscanning would break
8 : * repeatability, to say nothing of logically-irrelevant maintenance such
9 : * as physical extension or shortening of the relation.
10 : *
11 : * To achieve that, we proceed by hashing each candidate TID together with
12 : * the active seed, and then selecting it if the hash is less than the
13 : * cutoff value computed from the selection probability by BeginSampleScan.
14 : *
15 : *
16 : * Portions Copyright (c) 1996-2026, PostgreSQL Global Development Group
17 : * Portions Copyright (c) 1994, Regents of the University of California
18 : *
19 : * IDENTIFICATION
20 : * src/backend/access/tablesample/bernoulli.c
21 : *
22 : *-------------------------------------------------------------------------
23 : */
24 :
25 : #include "postgres.h"
26 :
27 : #include <math.h>
28 :
29 : #include "access/tsmapi.h"
30 : #include "catalog/pg_type.h"
31 : #include "common/hashfn.h"
32 : #include "optimizer/optimizer.h"
33 : #include "utils/fmgrprotos.h"
34 :
35 :
36 : /* Private state */
37 : typedef struct
38 : {
39 : uint64 cutoff; /* select tuples with hash less than this */
40 : uint32 seed; /* random seed */
41 : OffsetNumber lt; /* last tuple returned from current block */
42 : } BernoulliSamplerData;
43 :
44 :
45 : static void bernoulli_samplescangetsamplesize(PlannerInfo *root,
46 : RelOptInfo *baserel,
47 : List *paramexprs,
48 : BlockNumber *pages,
49 : double *tuples);
50 : static void bernoulli_initsamplescan(SampleScanState *node,
51 : int eflags);
52 : static void bernoulli_beginsamplescan(SampleScanState *node,
53 : Datum *params,
54 : int nparams,
55 : uint32 seed);
56 : static OffsetNumber bernoulli_nextsampletuple(SampleScanState *node,
57 : BlockNumber blockno,
58 : OffsetNumber maxoffset);
59 :
60 :
61 : /*
62 : * Create a TsmRoutine descriptor for the BERNOULLI method.
63 : */
64 : Datum
65 228 : tsm_bernoulli_handler(PG_FUNCTION_ARGS)
66 : {
67 228 : TsmRoutine *tsm = makeNode(TsmRoutine);
68 :
69 228 : tsm->parameterTypes = list_make1_oid(FLOAT4OID);
70 228 : tsm->repeatable_across_queries = true;
71 228 : tsm->repeatable_across_scans = true;
72 228 : tsm->SampleScanGetSampleSize = bernoulli_samplescangetsamplesize;
73 228 : tsm->InitSampleScan = bernoulli_initsamplescan;
74 228 : tsm->BeginSampleScan = bernoulli_beginsamplescan;
75 228 : tsm->NextSampleBlock = NULL;
76 228 : tsm->NextSampleTuple = bernoulli_nextsampletuple;
77 228 : tsm->EndSampleScan = NULL;
78 :
79 228 : PG_RETURN_POINTER(tsm);
80 : }
81 :
82 : /*
83 : * Sample size estimation.
84 : */
85 : static void
86 60 : bernoulli_samplescangetsamplesize(PlannerInfo *root,
87 : RelOptInfo *baserel,
88 : List *paramexprs,
89 : BlockNumber *pages,
90 : double *tuples)
91 : {
92 : Node *pctnode;
93 : float4 samplefract;
94 :
95 : /* Try to extract an estimate for the sample percentage */
96 60 : pctnode = (Node *) linitial(paramexprs);
97 60 : pctnode = estimate_expression_value(root, pctnode);
98 :
99 60 : if (IsA(pctnode, Const) &&
100 51 : !((Const *) pctnode)->constisnull)
101 : {
102 51 : samplefract = DatumGetFloat4(((Const *) pctnode)->constvalue);
103 51 : if (samplefract >= 0 && samplefract <= 100 && !isnan(samplefract))
104 45 : samplefract /= 100.0f;
105 : else
106 : {
107 : /* Default samplefract if the value is bogus */
108 6 : samplefract = 0.1f;
109 : }
110 : }
111 : else
112 : {
113 : /* Default samplefract if we didn't obtain a non-null Const */
114 9 : samplefract = 0.1f;
115 : }
116 :
117 : /* We'll visit all pages of the baserel */
118 60 : *pages = baserel->pages;
119 :
120 60 : *tuples = clamp_row_est(baserel->tuples * samplefract);
121 60 : }
122 :
123 : /*
124 : * Initialize during executor setup.
125 : */
126 : static void
127 60 : bernoulli_initsamplescan(SampleScanState *node, int eflags)
128 : {
129 60 : node->tsm_state = palloc0(sizeof(BernoulliSamplerData));
130 60 : }
131 :
132 : /*
133 : * Examine parameters and prepare for a sample scan.
134 : */
135 : static void
136 45 : bernoulli_beginsamplescan(SampleScanState *node,
137 : Datum *params,
138 : int nparams,
139 : uint32 seed)
140 : {
141 45 : BernoulliSamplerData *sampler = (BernoulliSamplerData *) node->tsm_state;
142 45 : double percent = DatumGetFloat4(params[0]);
143 : double dcutoff;
144 :
145 45 : if (percent < 0 || percent > 100 || isnan(percent))
146 6 : ereport(ERROR,
147 : (errcode(ERRCODE_INVALID_TABLESAMPLE_ARGUMENT),
148 : errmsg("sample percentage must be between 0 and 100")));
149 :
150 : /*
151 : * The cutoff is sample probability times (PG_UINT32_MAX + 1); we have to
152 : * store that as a uint64, of course. Note that this gives strictly
153 : * correct behavior at the limits of zero or one probability.
154 : */
155 39 : dcutoff = rint(((double) PG_UINT32_MAX + 1) * percent / 100);
156 39 : sampler->cutoff = (uint64) dcutoff;
157 39 : sampler->seed = seed;
158 39 : sampler->lt = InvalidOffsetNumber;
159 :
160 : /*
161 : * Use bulkread, since we're scanning all pages. But pagemode visibility
162 : * checking is a win only at larger sampling fractions. The 25% cutoff
163 : * here is based on very limited experimentation.
164 : */
165 39 : node->use_bulkread = true;
166 39 : node->use_pagemode = (percent >= 25);
167 39 : }
168 :
169 : /*
170 : * Select next sampled tuple in current block.
171 : *
172 : * It is OK here to return an offset without knowing if the tuple is visible
173 : * (or even exists). The reason is that we do the coinflip for every tuple
174 : * offset in the table. Since all tuples have the same probability of being
175 : * returned, it doesn't matter if we do extra coinflips for invisible tuples.
176 : *
177 : * When we reach end of the block, return InvalidOffsetNumber which tells
178 : * SampleScan to go to next block.
179 : */
180 : static OffsetNumber
181 64416 : bernoulli_nextsampletuple(SampleScanState *node,
182 : BlockNumber blockno,
183 : OffsetNumber maxoffset)
184 : {
185 64416 : BernoulliSamplerData *sampler = (BernoulliSamplerData *) node->tsm_state;
186 64416 : OffsetNumber tupoffset = sampler->lt;
187 : uint32 hashinput[3];
188 :
189 : /* Advance to first/next tuple in block */
190 64416 : if (tupoffset == InvalidOffsetNumber)
191 4194 : tupoffset = FirstOffsetNumber;
192 : else
193 60222 : tupoffset++;
194 :
195 : /*
196 : * We compute the hash by applying hash_any to an array of 3 uint32's
197 : * containing the block, offset, and seed. This is efficient to set up,
198 : * and with the current implementation of hash_any, it gives
199 : * machine-independent results, which is a nice property for regression
200 : * testing.
201 : *
202 : * These words in the hash input are the same throughout the block:
203 : */
204 64416 : hashinput[0] = blockno;
205 64416 : hashinput[2] = sampler->seed;
206 :
207 : /*
208 : * Loop over tuple offsets until finding suitable TID or reaching end of
209 : * block.
210 : */
211 124518 : for (; tupoffset <= maxoffset; tupoffset++)
212 : {
213 : uint32 hash;
214 :
215 120324 : hashinput[1] = tupoffset;
216 :
217 120324 : hash = DatumGetUInt32(hash_any((const unsigned char *) hashinput,
218 : (int) sizeof(hashinput)));
219 120324 : if (hash < sampler->cutoff)
220 60222 : break;
221 : }
222 :
223 64416 : if (tupoffset > maxoffset)
224 4194 : tupoffset = InvalidOffsetNumber;
225 :
226 64416 : sampler->lt = tupoffset;
227 :
228 64416 : return tupoffset;
229 : }
|
