Branch data Line data Source code
1 : : /*-------------------------------------------------------------------------
2 : : *
3 : : * file_fdw.c
4 : : * foreign-data wrapper for server-side flat files (or programs).
5 : : *
6 : : * Copyright (c) 2010-2026, PostgreSQL Global Development Group
7 : : *
8 : : * IDENTIFICATION
9 : : * contrib/file_fdw/file_fdw.c
10 : : *
11 : : *-------------------------------------------------------------------------
12 : : */
13 : : #include "postgres.h"
14 : :
15 : : #include <sys/stat.h>
16 : : #include <unistd.h>
17 : :
18 : : #include "access/htup_details.h"
19 : : #include "access/reloptions.h"
20 : : #include "access/sysattr.h"
21 : : #include "access/table.h"
22 : : #include "catalog/pg_authid.h"
23 : : #include "catalog/pg_foreign_table.h"
24 : : #include "commands/copy.h"
25 : : #include "commands/copyfrom_internal.h"
26 : : #include "commands/defrem.h"
27 : : #include "commands/explain_format.h"
28 : : #include "commands/explain_state.h"
29 : : #include "commands/vacuum.h"
30 : : #include "executor/executor.h"
31 : : #include "foreign/fdwapi.h"
32 : : #include "foreign/foreign.h"
33 : : #include "miscadmin.h"
34 : : #include "nodes/makefuncs.h"
35 : : #include "optimizer/optimizer.h"
36 : : #include "optimizer/pathnode.h"
37 : : #include "optimizer/planmain.h"
38 : : #include "optimizer/restrictinfo.h"
39 : : #include "utils/acl.h"
40 : : #include "utils/memutils.h"
41 : : #include "utils/rel.h"
42 : : #include "utils/sampling.h"
43 : : #include "utils/varlena.h"
44 : :
45 : 1 : PG_MODULE_MAGIC_EXT(
46 : : .name = "file_fdw",
47 : : .version = PG_VERSION
48 : : );
49 : :
50 : : /*
51 : : * Describes the valid options for objects that use this wrapper.
52 : : */
53 : : struct FileFdwOption
54 : : {
55 : : const char *optname;
56 : : Oid optcontext; /* Oid of catalog in which option may appear */
57 : : };
58 : :
59 : : /*
60 : : * Valid options for file_fdw.
61 : : * These options are based on the options for the COPY FROM command.
62 : : * But note that force_not_null and force_null are handled as boolean options
63 : : * attached to a column, not as table options.
64 : : *
65 : : * Note: If you are adding new option for user mapping, you need to modify
66 : : * fileGetOptions(), which currently doesn't bother to look at user mappings.
67 : : */
68 : : static const struct FileFdwOption valid_options[] = {
69 : : /* Data source options */
70 : : {"filename", ForeignTableRelationId},
71 : : {"program", ForeignTableRelationId},
72 : :
73 : : /* Format options */
74 : : /* oids option is not supported */
75 : : {"format", ForeignTableRelationId},
76 : : {"header", ForeignTableRelationId},
77 : : {"delimiter", ForeignTableRelationId},
78 : : {"quote", ForeignTableRelationId},
79 : : {"escape", ForeignTableRelationId},
80 : : {"null", ForeignTableRelationId},
81 : : {"default", ForeignTableRelationId},
82 : : {"encoding", ForeignTableRelationId},
83 : : {"on_error", ForeignTableRelationId},
84 : : {"log_verbosity", ForeignTableRelationId},
85 : : {"reject_limit", ForeignTableRelationId},
86 : : {"force_not_null", AttributeRelationId},
87 : : {"force_null", AttributeRelationId},
88 : :
89 : : /*
90 : : * force_quote is not supported by file_fdw because it's for COPY TO.
91 : : */
92 : :
93 : : /* Sentinel */
94 : : {NULL, InvalidOid}
95 : : };
96 : :
97 : : /*
98 : : * FDW-specific information for RelOptInfo.fdw_private.
99 : : */
100 : : typedef struct FileFdwPlanState
101 : : {
102 : : char *filename; /* file or program to read from */
103 : : bool is_program; /* true if filename represents an OS command */
104 : : List *options; /* merged COPY options, excluding filename and
105 : : * is_program */
106 : : BlockNumber pages; /* estimate of file's physical size */
107 : : double ntuples; /* estimate of number of data rows */
108 : : } FileFdwPlanState;
109 : :
110 : : /*
111 : : * FDW-specific information for ForeignScanState.fdw_state.
112 : : */
113 : : typedef struct FileFdwExecutionState
114 : : {
115 : : char *filename; /* file or program to read from */
116 : : bool is_program; /* true if filename represents an OS command */
117 : : List *options; /* merged COPY options, excluding filename and
118 : : * is_program */
119 : : CopyFromState cstate; /* COPY execution state */
120 : : } FileFdwExecutionState;
121 : :
122 : : /*
123 : : * SQL functions
124 : : */
125 : 2 : PG_FUNCTION_INFO_V1(file_fdw_handler);
126 : 2 : PG_FUNCTION_INFO_V1(file_fdw_validator);
127 : :
128 : : /*
129 : : * FDW callback routines
130 : : */
131 : : static void fileGetForeignRelSize(PlannerInfo *root,
132 : : RelOptInfo *baserel,
133 : : Oid foreigntableid);
134 : : static void fileGetForeignPaths(PlannerInfo *root,
135 : : RelOptInfo *baserel,
136 : : Oid foreigntableid);
137 : : static ForeignScan *fileGetForeignPlan(PlannerInfo *root,
138 : : RelOptInfo *baserel,
139 : : Oid foreigntableid,
140 : : ForeignPath *best_path,
141 : : List *tlist,
142 : : List *scan_clauses,
143 : : Plan *outer_plan);
144 : : static void fileExplainForeignScan(ForeignScanState *node, ExplainState *es);
145 : : static void fileBeginForeignScan(ForeignScanState *node, int eflags);
146 : : static TupleTableSlot *fileIterateForeignScan(ForeignScanState *node);
147 : : static void fileReScanForeignScan(ForeignScanState *node);
148 : : static void fileEndForeignScan(ForeignScanState *node);
149 : : static bool fileAnalyzeForeignTable(Relation relation,
150 : : AcquireSampleRowsFunc *func,
151 : : BlockNumber *totalpages);
152 : : static bool fileIsForeignScanParallelSafe(PlannerInfo *root, RelOptInfo *rel,
153 : : RangeTblEntry *rte);
154 : :
155 : : /*
156 : : * Helper functions
157 : : */
158 : : static bool is_valid_option(const char *option, Oid context);
159 : : static void fileGetOptions(Oid foreigntableid,
160 : : char **filename,
161 : : bool *is_program,
162 : : List **other_options);
163 : : static List *get_file_fdw_attribute_options(Oid relid);
164 : : static bool check_selective_binary_conversion(RelOptInfo *baserel,
165 : : Oid foreigntableid,
166 : : List **columns);
167 : : static void estimate_size(PlannerInfo *root, RelOptInfo *baserel,
168 : : FileFdwPlanState *fdw_private);
169 : : static void estimate_costs(PlannerInfo *root, RelOptInfo *baserel,
170 : : FileFdwPlanState *fdw_private,
171 : : Cost *startup_cost, Cost *total_cost);
172 : : static int file_acquire_sample_rows(Relation onerel, int elevel,
173 : : HeapTuple *rows, int targrows,
174 : : double *totalrows, double *totaldeadrows);
175 : :
176 : :
177 : : /*
178 : : * Foreign-data wrapper handler function: return a struct with pointers
179 : : * to my callback routines.
180 : : */
181 : : Datum
182 : 22 : file_fdw_handler(PG_FUNCTION_ARGS)
183 : : {
184 : 22 : FdwRoutine *fdwroutine = makeNode(FdwRoutine);
185 : :
186 : 22 : fdwroutine->GetForeignRelSize = fileGetForeignRelSize;
187 : 22 : fdwroutine->GetForeignPaths = fileGetForeignPaths;
188 : 22 : fdwroutine->GetForeignPlan = fileGetForeignPlan;
189 : 22 : fdwroutine->ExplainForeignScan = fileExplainForeignScan;
190 : 22 : fdwroutine->BeginForeignScan = fileBeginForeignScan;
191 : 22 : fdwroutine->IterateForeignScan = fileIterateForeignScan;
192 : 22 : fdwroutine->ReScanForeignScan = fileReScanForeignScan;
193 : 22 : fdwroutine->EndForeignScan = fileEndForeignScan;
194 : 22 : fdwroutine->AnalyzeForeignTable = fileAnalyzeForeignTable;
195 : 22 : fdwroutine->IsForeignScanParallelSafe = fileIsForeignScanParallelSafe;
196 : :
197 : 22 : PG_RETURN_POINTER(fdwroutine);
198 : : }
199 : :
200 : : /*
201 : : * Validate the generic options given to a FOREIGN DATA WRAPPER, SERVER,
202 : : * USER MAPPING or FOREIGN TABLE that uses file_fdw.
203 : : *
204 : : * Raise an ERROR if the option or its value is considered invalid.
205 : : */
206 : : Datum
207 : 68 : file_fdw_validator(PG_FUNCTION_ARGS)
208 : : {
209 : 68 : List *options_list = untransformRelOptions(PG_GETARG_DATUM(0));
210 : 68 : Oid catalog = PG_GETARG_OID(1);
211 : 68 : char *filename = NULL;
212 : 68 : DefElem *force_not_null = NULL;
213 : 68 : DefElem *force_null = NULL;
214 : 68 : List *other_options = NIL;
215 : : ListCell *cell;
216 : :
217 : : /*
218 : : * Check that only options supported by file_fdw, and allowed for the
219 : : * current object type, are given.
220 : : */
221 [ + + + + : 225 : foreach(cell, options_list)
+ + ]
222 : : {
223 : 167 : DefElem *def = (DefElem *) lfirst(cell);
224 : :
225 [ + + ]: 167 : if (!is_valid_option(def->defname, catalog))
226 : : {
227 : : const struct FileFdwOption *opt;
228 : : const char *closest_match;
229 : : ClosestMatchState match_state;
230 : 9 : bool has_valid_options = false;
231 : :
232 : : /*
233 : : * Unknown option specified, complain about it. Provide a hint
234 : : * with a valid option that looks similar, if there is one.
235 : : */
236 : 9 : initClosestMatch(&match_state, def->defname, 4);
237 [ + + ]: 144 : for (opt = valid_options; opt->optname; opt++)
238 : : {
239 [ + + ]: 135 : if (catalog == opt->optcontext)
240 : : {
241 : 39 : has_valid_options = true;
242 : 39 : updateClosestMatch(&match_state, opt->optname);
243 : : }
244 : : }
245 : :
246 : 9 : closest_match = getClosestMatch(&match_state);
247 [ + - + + : 9 : ereport(ERROR,
- + ]
248 : : (errcode(ERRCODE_FDW_INVALID_OPTION_NAME),
249 : : errmsg("invalid option \"%s\"", def->defname),
250 : : has_valid_options ? closest_match ?
251 : : errhint("Perhaps you meant the option \"%s\".",
252 : : closest_match) : 0 :
253 : : errhint("There are no valid options in this context.")));
254 : : }
255 : :
256 : : /*
257 : : * Separate out filename, program, and column-specific options, since
258 : : * ProcessCopyOptions won't accept them.
259 : : */
260 [ + + ]: 158 : if (strcmp(def->defname, "filename") == 0 ||
261 [ - + ]: 139 : strcmp(def->defname, "program") == 0)
262 : : {
263 [ - + ]: 19 : if (filename)
264 [ # # ]: 0 : ereport(ERROR,
265 : : (errcode(ERRCODE_SYNTAX_ERROR),
266 : : errmsg("conflicting or redundant options")));
267 : :
268 : : /*
269 : : * Check permissions for changing which file or program is used by
270 : : * the file_fdw.
271 : : *
272 : : * Only members of the role 'pg_read_server_files' are allowed to
273 : : * set the 'filename' option of a file_fdw foreign table, while
274 : : * only members of the role 'pg_execute_server_program' are
275 : : * allowed to set the 'program' option. This is because we don't
276 : : * want regular users to be able to control which file gets read
277 : : * or which program gets executed.
278 : : *
279 : : * Putting this sort of permissions check in a validator is a bit
280 : : * of a crock, but there doesn't seem to be any other place that
281 : : * can enforce the check more cleanly.
282 : : *
283 : : * Note that the valid_options[] array disallows setting filename
284 : : * and program at any options level other than foreign table ---
285 : : * otherwise there'd still be a security hole.
286 : : */
287 [ + - ]: 19 : if (strcmp(def->defname, "filename") == 0 &&
288 [ + + ]: 19 : !has_privs_of_role(GetUserId(), ROLE_PG_READ_SERVER_FILES))
289 [ + - ]: 1 : ereport(ERROR,
290 : : (errcode(ERRCODE_INSUFFICIENT_PRIVILEGE),
291 : : errmsg("permission denied to set the \"%s\" option of a file_fdw foreign table",
292 : : "filename"),
293 : : errdetail("Only roles with privileges of the \"%s\" role may set this option.",
294 : : "pg_read_server_files")));
295 : :
296 [ - + ]: 18 : if (strcmp(def->defname, "program") == 0 &&
297 [ # # ]: 0 : !has_privs_of_role(GetUserId(), ROLE_PG_EXECUTE_SERVER_PROGRAM))
298 [ # # ]: 0 : ereport(ERROR,
299 : : (errcode(ERRCODE_INSUFFICIENT_PRIVILEGE),
300 : : errmsg("permission denied to set the \"%s\" option of a file_fdw foreign table",
301 : : "program"),
302 : : errdetail("Only roles with privileges of the \"%s\" role may set this option.",
303 : : "pg_execute_server_program")));
304 : :
305 : 18 : filename = defGetString(def);
306 : : }
307 : :
308 : : /*
309 : : * force_not_null is a boolean option; after validation we can discard
310 : : * it - it will be retrieved later in get_file_fdw_attribute_options()
311 : : */
312 [ + + ]: 139 : else if (strcmp(def->defname, "force_not_null") == 0)
313 : : {
314 [ - + ]: 4 : if (force_not_null)
315 [ # # ]: 0 : ereport(ERROR,
316 : : (errcode(ERRCODE_SYNTAX_ERROR),
317 : : errmsg("conflicting or redundant options"),
318 : : errhint("Option \"force_not_null\" supplied more than once for a column.")));
319 : 4 : force_not_null = def;
320 : : /* Don't care what the value is, as long as it's a legal boolean */
321 : 4 : (void) defGetBoolean(def);
322 : : }
323 : : /* See comments for force_not_null above */
324 [ + + ]: 135 : else if (strcmp(def->defname, "force_null") == 0)
325 : : {
326 [ - + ]: 4 : if (force_null)
327 [ # # ]: 0 : ereport(ERROR,
328 : : (errcode(ERRCODE_SYNTAX_ERROR),
329 : : errmsg("conflicting or redundant options"),
330 : : errhint("Option \"force_null\" supplied more than once for a column.")));
331 : 4 : force_null = def;
332 : 4 : (void) defGetBoolean(def);
333 : : }
334 : : else
335 : 131 : other_options = lappend(other_options, def);
336 : : }
337 : :
338 : : /*
339 : : * Now apply the core COPY code's validation logic for more checks.
340 : : */
341 : 58 : ProcessCopyOptions(NULL, NULL, true, other_options);
342 : :
343 : : /*
344 : : * Either filename or program option is required for file_fdw foreign
345 : : * tables.
346 : : */
347 [ + + + + ]: 30 : if (catalog == ForeignTableRelationId && filename == NULL)
348 [ + - ]: 1 : ereport(ERROR,
349 : : (errcode(ERRCODE_FDW_DYNAMIC_PARAMETER_VALUE_NEEDED),
350 : : errmsg("either filename or program is required for file_fdw foreign tables")));
351 : :
352 : 29 : PG_RETURN_VOID();
353 : : }
354 : :
355 : : /*
356 : : * Check if the provided option is one of the valid options.
357 : : * context is the Oid of the catalog holding the object the option is for.
358 : : */
359 : : static bool
360 : 167 : is_valid_option(const char *option, Oid context)
361 : : {
362 : : const struct FileFdwOption *opt;
363 : :
364 [ + + ]: 1027 : for (opt = valid_options; opt->optname; opt++)
365 : : {
366 [ + + + + ]: 1018 : if (context == opt->optcontext && strcmp(opt->optname, option) == 0)
367 : 158 : return true;
368 : : }
369 : 9 : return false;
370 : : }
371 : :
372 : : /*
373 : : * Fetch the options for a file_fdw foreign table.
374 : : *
375 : : * We have to separate out filename/program from the other options because
376 : : * those must not appear in the options list passed to the core COPY code.
377 : : */
378 : : static void
379 : 83 : fileGetOptions(Oid foreigntableid,
380 : : char **filename, bool *is_program, List **other_options)
381 : : {
382 : : ForeignTable *table;
383 : : ForeignServer *server;
384 : : ForeignDataWrapper *wrapper;
385 : : List *options;
386 : : ListCell *lc;
387 : :
388 : : /*
389 : : * Extract options from FDW objects. We ignore user mappings because
390 : : * file_fdw doesn't have any options that can be specified there.
391 : : *
392 : : * (XXX Actually, given the current contents of valid_options[], there's
393 : : * no point in examining anything except the foreign table's own options.
394 : : * Simplify?)
395 : : */
396 : 83 : table = GetForeignTable(foreigntableid);
397 : 83 : server = GetForeignServer(table->serverid);
398 : 83 : wrapper = GetForeignDataWrapper(server->fdwid);
399 : :
400 : 83 : options = NIL;
401 : 83 : options = list_concat(options, wrapper->options);
402 : 83 : options = list_concat(options, server->options);
403 : 83 : options = list_concat(options, table->options);
404 : 83 : options = list_concat(options, get_file_fdw_attribute_options(foreigntableid));
405 : :
406 : : /*
407 : : * Separate out the filename or program option (we assume there is only
408 : : * one).
409 : : */
410 : 83 : *filename = NULL;
411 : 83 : *is_program = false;
412 [ + - + - : 166 : foreach(lc, options)
+ - ]
413 : : {
414 : 166 : DefElem *def = (DefElem *) lfirst(lc);
415 : :
416 [ + + ]: 166 : if (strcmp(def->defname, "filename") == 0)
417 : : {
418 : 83 : *filename = defGetString(def);
419 : 83 : options = foreach_delete_current(options, lc);
420 : 83 : break;
421 : : }
422 [ - + ]: 83 : else if (strcmp(def->defname, "program") == 0)
423 : : {
424 : 0 : *filename = defGetString(def);
425 : 0 : *is_program = true;
426 : 0 : options = foreach_delete_current(options, lc);
427 : 0 : break;
428 : : }
429 : : }
430 : :
431 : : /*
432 : : * The validator should have checked that filename or program was included
433 : : * in the options, but check again, just in case.
434 : : */
435 [ - + ]: 83 : if (*filename == NULL)
436 [ # # ]: 0 : elog(ERROR, "either filename or program is required for file_fdw foreign tables");
437 : :
438 : 83 : *other_options = options;
439 : 83 : }
440 : :
441 : : /*
442 : : * Retrieve per-column generic options from pg_attribute and construct a list
443 : : * of DefElems representing them.
444 : : *
445 : : * At the moment we only have "force_not_null", and "force_null",
446 : : * which should each be combined into a single DefElem listing all such
447 : : * columns, since that's what COPY expects.
448 : : */
449 : : static List *
450 : 83 : get_file_fdw_attribute_options(Oid relid)
451 : : {
452 : : Relation rel;
453 : : TupleDesc tupleDesc;
454 : : AttrNumber natts;
455 : : AttrNumber attnum;
456 : 83 : List *fnncolumns = NIL;
457 : 83 : List *fncolumns = NIL;
458 : :
459 : 83 : List *options = NIL;
460 : :
461 : 83 : rel = table_open(relid, AccessShareLock);
462 : 83 : tupleDesc = RelationGetDescr(rel);
463 : 83 : natts = tupleDesc->natts;
464 : :
465 : : /* Retrieve FDW options for all user-defined attributes. */
466 [ + + ]: 261 : for (attnum = 1; attnum <= natts; attnum++)
467 : : {
468 : 178 : Form_pg_attribute attr = TupleDescAttr(tupleDesc, attnum - 1);
469 : : List *column_options;
470 : : ListCell *lc;
471 : :
472 : : /* Skip dropped attributes. */
473 [ - + ]: 178 : if (attr->attisdropped)
474 : 0 : continue;
475 : :
476 : 178 : column_options = GetForeignColumnOptions(relid, attnum);
477 [ + + + + : 194 : foreach(lc, column_options)
+ + ]
478 : : {
479 : 16 : DefElem *def = (DefElem *) lfirst(lc);
480 : :
481 [ + + ]: 16 : if (strcmp(def->defname, "force_not_null") == 0)
482 : : {
483 [ + + ]: 8 : if (defGetBoolean(def))
484 : : {
485 : 4 : char *attname = pstrdup(NameStr(attr->attname));
486 : :
487 : 4 : fnncolumns = lappend(fnncolumns, makeString(attname));
488 : : }
489 : : }
490 [ + - ]: 8 : else if (strcmp(def->defname, "force_null") == 0)
491 : : {
492 [ + + ]: 8 : if (defGetBoolean(def))
493 : : {
494 : 4 : char *attname = pstrdup(NameStr(attr->attname));
495 : :
496 : 4 : fncolumns = lappend(fncolumns, makeString(attname));
497 : : }
498 : : }
499 : : /* maybe in future handle other column options here */
500 : : }
501 : : }
502 : :
503 : 83 : table_close(rel, AccessShareLock);
504 : :
505 : : /*
506 : : * Return DefElem only when some column(s) have force_not_null /
507 : : * force_null options set
508 : : */
509 [ + + ]: 83 : if (fnncolumns != NIL)
510 : 4 : options = lappend(options, makeDefElem("force_not_null", (Node *) fnncolumns, -1));
511 : :
512 [ + + ]: 83 : if (fncolumns != NIL)
513 : 4 : options = lappend(options, makeDefElem("force_null", (Node *) fncolumns, -1));
514 : :
515 : 83 : return options;
516 : : }
517 : :
518 : : /*
519 : : * fileGetForeignRelSize
520 : : * Obtain relation size estimates for a foreign table
521 : : */
522 : : static void
523 : 43 : fileGetForeignRelSize(PlannerInfo *root,
524 : : RelOptInfo *baserel,
525 : : Oid foreigntableid)
526 : : {
527 : : FileFdwPlanState *fdw_private;
528 : :
529 : : /*
530 : : * Fetch options. We only need filename (or program) at this point, but
531 : : * we might as well get everything and not need to re-fetch it later in
532 : : * planning.
533 : : */
534 : 43 : fdw_private = palloc_object(FileFdwPlanState);
535 : 43 : fileGetOptions(foreigntableid,
536 : : &fdw_private->filename,
537 : : &fdw_private->is_program,
538 : : &fdw_private->options);
539 : 43 : baserel->fdw_private = fdw_private;
540 : :
541 : : /* Estimate relation size */
542 : 43 : estimate_size(root, baserel, fdw_private);
543 : 43 : }
544 : :
545 : : /*
546 : : * fileGetForeignPaths
547 : : * Create possible access paths for a scan on the foreign table
548 : : *
549 : : * Currently we don't support any push-down feature, so there is only one
550 : : * possible access path, which simply returns all records in the order in
551 : : * the data file.
552 : : */
553 : : static void
554 : 43 : fileGetForeignPaths(PlannerInfo *root,
555 : : RelOptInfo *baserel,
556 : : Oid foreigntableid)
557 : : {
558 : 43 : FileFdwPlanState *fdw_private = (FileFdwPlanState *) baserel->fdw_private;
559 : : Cost startup_cost;
560 : : Cost total_cost;
561 : : List *columns;
562 : 43 : List *coptions = NIL;
563 : :
564 : : /* Decide whether to selectively perform binary conversion */
565 [ + + ]: 43 : if (check_selective_binary_conversion(baserel,
566 : : foreigntableid,
567 : : &columns))
568 : 5 : coptions = list_make1(makeDefElem("convert_selectively",
569 : : (Node *) columns, -1));
570 : :
571 : : /* Estimate costs */
572 : 43 : estimate_costs(root, baserel, fdw_private,
573 : : &startup_cost, &total_cost);
574 : :
575 : : /*
576 : : * Create a ForeignPath node and add it as only possible path. We use the
577 : : * fdw_private list of the path to carry the convert_selectively option;
578 : : * it will be propagated into the fdw_private list of the Plan node.
579 : : *
580 : : * We don't support pushing join clauses into the quals of this path, but
581 : : * it could still have required parameterization due to LATERAL refs in
582 : : * its tlist.
583 : : */
584 : 43 : add_path(baserel, (Path *)
585 : 43 : create_foreignscan_path(root, baserel,
586 : : NULL, /* default pathtarget */
587 : : baserel->rows,
588 : : 0,
589 : : startup_cost,
590 : : total_cost,
591 : : NIL, /* no pathkeys */
592 : : baserel->lateral_relids,
593 : : NULL, /* no extra plan */
594 : : NIL, /* no fdw_restrictinfo list */
595 : : coptions));
596 : :
597 : : /*
598 : : * If data file was sorted, and we knew it somehow, we could insert
599 : : * appropriate pathkeys into the ForeignPath node to tell the planner
600 : : * that.
601 : : */
602 : 43 : }
603 : :
604 : : /*
605 : : * fileGetForeignPlan
606 : : * Create a ForeignScan plan node for scanning the foreign table
607 : : */
608 : : static ForeignScan *
609 : 43 : fileGetForeignPlan(PlannerInfo *root,
610 : : RelOptInfo *baserel,
611 : : Oid foreigntableid,
612 : : ForeignPath *best_path,
613 : : List *tlist,
614 : : List *scan_clauses,
615 : : Plan *outer_plan)
616 : : {
617 : 43 : Index scan_relid = baserel->relid;
618 : :
619 : : /*
620 : : * We have no native ability to evaluate restriction clauses, so we just
621 : : * put all the scan_clauses into the plan node's qual list for the
622 : : * executor to check. So all we have to do here is strip RestrictInfo
623 : : * nodes from the clauses and ignore pseudoconstants (which will be
624 : : * handled elsewhere).
625 : : */
626 : 43 : scan_clauses = extract_actual_clauses(scan_clauses, false);
627 : :
628 : : /* Create the ForeignScan node */
629 : 43 : return make_foreignscan(tlist,
630 : : scan_clauses,
631 : : scan_relid,
632 : : NIL, /* no expressions to evaluate */
633 : : best_path->fdw_private,
634 : : NIL, /* no custom tlist */
635 : : NIL, /* no remote quals */
636 : : outer_plan);
637 : : }
638 : :
639 : : /*
640 : : * fileExplainForeignScan
641 : : * Produce extra output for EXPLAIN
642 : : */
643 : : static void
644 : 3 : fileExplainForeignScan(ForeignScanState *node, ExplainState *es)
645 : : {
646 : : char *filename;
647 : : bool is_program;
648 : : List *options;
649 : :
650 : : /* Fetch options --- we only need filename and is_program at this point */
651 : 3 : fileGetOptions(RelationGetRelid(node->ss.ss_currentRelation),
652 : : &filename, &is_program, &options);
653 : :
654 [ - + ]: 3 : if (is_program)
655 : 0 : ExplainPropertyText("Foreign Program", filename, es);
656 : : else
657 : 3 : ExplainPropertyText("Foreign File", filename, es);
658 : :
659 : : /* Suppress file size if we're not showing cost details */
660 [ - + ]: 3 : if (es->costs)
661 : : {
662 : : struct stat stat_buf;
663 : :
664 [ # # # # ]: 0 : if (!is_program &&
665 : 0 : stat(filename, &stat_buf) == 0)
666 : 0 : ExplainPropertyInteger("Foreign File Size", "b",
667 : 0 : (int64) stat_buf.st_size, es);
668 : : }
669 : 3 : }
670 : :
671 : : /*
672 : : * fileBeginForeignScan
673 : : * Initiate access to the file by creating CopyState
674 : : */
675 : : static void
676 : 38 : fileBeginForeignScan(ForeignScanState *node, int eflags)
677 : : {
678 : 38 : ForeignScan *plan = (ForeignScan *) node->ss.ps.plan;
679 : : char *filename;
680 : : bool is_program;
681 : : List *options;
682 : : CopyFromState cstate;
683 : : FileFdwExecutionState *festate;
684 : :
685 : : /*
686 : : * Do nothing in EXPLAIN (no ANALYZE) case. node->fdw_state stays NULL.
687 : : */
688 [ + + ]: 38 : if (eflags & EXEC_FLAG_EXPLAIN_ONLY)
689 : 3 : return;
690 : :
691 : : /* Fetch options of foreign table */
692 : 35 : fileGetOptions(RelationGetRelid(node->ss.ss_currentRelation),
693 : : &filename, &is_program, &options);
694 : :
695 : : /* Add any options from the plan (currently only convert_selectively) */
696 : 35 : options = list_concat(options, plan->fdw_private);
697 : :
698 : : /*
699 : : * Create CopyState from FDW options. We always acquire all columns, so
700 : : * as to match the expected ScanTupleSlot signature.
701 : : */
702 : 35 : cstate = BeginCopyFrom(NULL,
703 : : node->ss.ss_currentRelation,
704 : : NULL,
705 : : filename,
706 : : is_program,
707 : : NULL,
708 : : NIL,
709 : : options);
710 : :
711 : : /*
712 : : * Save state in node->fdw_state. We must save enough information to call
713 : : * BeginCopyFrom() again.
714 : : */
715 : 34 : festate = palloc_object(FileFdwExecutionState);
716 : 34 : festate->filename = filename;
717 : 34 : festate->is_program = is_program;
718 : 34 : festate->options = options;
719 : 34 : festate->cstate = cstate;
720 : :
721 : 34 : node->fdw_state = festate;
722 : : }
723 : :
724 : : /*
725 : : * fileIterateForeignScan
726 : : * Read next record from the data file and store it into the
727 : : * ScanTupleSlot as a virtual tuple
728 : : */
729 : : static TupleTableSlot *
730 : 132 : fileIterateForeignScan(ForeignScanState *node)
731 : : {
732 : 132 : FileFdwExecutionState *festate = (FileFdwExecutionState *) node->fdw_state;
733 : 132 : EState *estate = CreateExecutorState();
734 : : ExprContext *econtext;
735 : 132 : MemoryContext oldcontext = CurrentMemoryContext;
736 : 132 : TupleTableSlot *slot = node->ss.ss_ScanTupleSlot;
737 : 132 : CopyFromState cstate = festate->cstate;
738 : : ErrorContextCallback errcallback;
739 : :
740 : : /* Set up callback to identify error line number. */
741 : 132 : errcallback.callback = CopyFromErrorCallback;
742 : 132 : errcallback.arg = cstate;
743 : 132 : errcallback.previous = error_context_stack;
744 : 132 : error_context_stack = &errcallback;
745 : :
746 : : /*
747 : : * We pass ExprContext because there might be a use of the DEFAULT option
748 : : * in COPY FROM, so we may need to evaluate default expressions.
749 : : */
750 [ - + ]: 132 : econtext = GetPerTupleExprContext(estate);
751 : :
752 : 139 : retry:
753 : :
754 : : /*
755 : : * DEFAULT expressions need to be evaluated in a per-tuple context, so
756 : : * switch in case we are doing that.
757 : : */
758 [ + - ]: 139 : MemoryContextSwitchTo(GetPerTupleMemoryContext(estate));
759 : :
760 : : /*
761 : : * The protocol for loading a virtual tuple into a slot is first
762 : : * ExecClearTuple, then fill the values/isnull arrays, then
763 : : * ExecStoreVirtualTuple. If we don't find another row in the file, we
764 : : * just skip the last step, leaving the slot empty as required.
765 : : *
766 : : */
767 : 139 : ExecClearTuple(slot);
768 : :
769 [ + + ]: 139 : if (NextCopyFrom(cstate, econtext, slot->tts_values, slot->tts_isnull))
770 : : {
771 [ + + ]: 104 : if (cstate->opts.on_error == COPY_ON_ERROR_IGNORE &&
772 [ + + ]: 16 : cstate->escontext->error_occurred)
773 : : {
774 : : /*
775 : : * Soft error occurred, skip this tuple and just make
776 : : * ErrorSaveContext ready for the next NextCopyFrom. Since we
777 : : * don't set details_wanted and error_data is not to be filled,
778 : : * just resetting error_occurred is enough.
779 : : */
780 : 8 : cstate->escontext->error_occurred = false;
781 : :
782 : : /* Switch back to original memory context */
783 : 8 : MemoryContextSwitchTo(oldcontext);
784 : :
785 : : /*
786 : : * Make sure we are interruptible while repeatedly calling
787 : : * NextCopyFrom() until no soft error occurs.
788 : : */
789 [ - + ]: 8 : CHECK_FOR_INTERRUPTS();
790 : :
791 : : /*
792 : : * Reset the per-tuple exprcontext, to clean-up after expression
793 : : * evaluations etc.
794 : : */
795 [ + - ]: 8 : ResetPerTupleExprContext(estate);
796 : :
797 [ + + ]: 8 : if (cstate->opts.reject_limit > 0 &&
798 [ + + ]: 4 : cstate->num_errors > cstate->opts.reject_limit)
799 [ + - ]: 1 : ereport(ERROR,
800 : : (errcode(ERRCODE_INVALID_TEXT_REPRESENTATION),
801 : : errmsg("skipped more than REJECT_LIMIT (%" PRId64 ") rows due to data type incompatibility",
802 : : cstate->opts.reject_limit)));
803 : :
804 : : /* Repeat NextCopyFrom() until no soft error occurs */
805 : 7 : goto retry;
806 : : }
807 : :
808 : 96 : ExecStoreVirtualTuple(slot);
809 : : }
810 : :
811 : : /* Switch back to original memory context */
812 : 129 : MemoryContextSwitchTo(oldcontext);
813 : :
814 : : /* Remove error callback. */
815 : 129 : error_context_stack = errcallback.previous;
816 : :
817 : 129 : return slot;
818 : : }
819 : :
820 : : /*
821 : : * fileReScanForeignScan
822 : : * Rescan table, possibly with new parameters
823 : : */
824 : : static void
825 : 3 : fileReScanForeignScan(ForeignScanState *node)
826 : : {
827 : 3 : FileFdwExecutionState *festate = (FileFdwExecutionState *) node->fdw_state;
828 : :
829 : 3 : EndCopyFrom(festate->cstate);
830 : :
831 : 6 : festate->cstate = BeginCopyFrom(NULL,
832 : : node->ss.ss_currentRelation,
833 : : NULL,
834 : 3 : festate->filename,
835 : 3 : festate->is_program,
836 : : NULL,
837 : : NIL,
838 : : festate->options);
839 : 3 : }
840 : :
841 : : /*
842 : : * fileEndForeignScan
843 : : * Finish scanning foreign table and dispose objects used for this scan
844 : : */
845 : : static void
846 : 34 : fileEndForeignScan(ForeignScanState *node)
847 : : {
848 : 34 : FileFdwExecutionState *festate = (FileFdwExecutionState *) node->fdw_state;
849 : :
850 : : /* if festate is NULL, we are in EXPLAIN; nothing to do */
851 [ + + ]: 34 : if (!festate)
852 : 3 : return;
853 : :
854 [ + + ]: 31 : if (festate->cstate->opts.on_error == COPY_ON_ERROR_IGNORE &&
855 [ + - ]: 3 : festate->cstate->num_errors > 0 &&
856 [ + + ]: 3 : festate->cstate->opts.log_verbosity >= COPY_LOG_VERBOSITY_DEFAULT)
857 [ + - ]: 1 : ereport(NOTICE,
858 : : errmsg_plural("%" PRIu64 " row was skipped due to data type incompatibility",
859 : : "%" PRIu64 " rows were skipped due to data type incompatibility",
860 : : festate->cstate->num_errors,
861 : : festate->cstate->num_errors));
862 : :
863 : 31 : EndCopyFrom(festate->cstate);
864 : : }
865 : :
866 : : /*
867 : : * fileAnalyzeForeignTable
868 : : * Test whether analyzing this foreign table is supported
869 : : */
870 : : static bool
871 : 1 : fileAnalyzeForeignTable(Relation relation,
872 : : AcquireSampleRowsFunc *func,
873 : : BlockNumber *totalpages)
874 : : {
875 : : char *filename;
876 : : bool is_program;
877 : : List *options;
878 : : struct stat stat_buf;
879 : :
880 : : /* Fetch options of foreign table */
881 : 1 : fileGetOptions(RelationGetRelid(relation), &filename, &is_program, &options);
882 : :
883 : : /*
884 : : * If this is a program instead of a file, just return false to skip
885 : : * analyzing the table. We could run the program and collect stats on
886 : : * whatever it currently returns, but it seems likely that in such cases
887 : : * the output would be too volatile for the stats to be useful. Maybe
888 : : * there should be an option to enable doing this?
889 : : */
890 [ - + ]: 1 : if (is_program)
891 : 0 : return false;
892 : :
893 : : /*
894 : : * Get size of the file. (XXX if we fail here, would it be better to just
895 : : * return false to skip analyzing the table?)
896 : : */
897 [ - + ]: 1 : if (stat(filename, &stat_buf) < 0)
898 [ # # ]: 0 : ereport(ERROR,
899 : : (errcode_for_file_access(),
900 : : errmsg("could not stat file \"%s\": %m",
901 : : filename)));
902 : :
903 : : /*
904 : : * Convert size to pages. Must return at least 1 so that we can tell
905 : : * later on that pg_class.relpages is not default.
906 : : */
907 : 1 : *totalpages = (stat_buf.st_size + (BLCKSZ - 1)) / BLCKSZ;
908 [ - + ]: 1 : if (*totalpages < 1)
909 : 0 : *totalpages = 1;
910 : :
911 : 1 : *func = file_acquire_sample_rows;
912 : :
913 : 1 : return true;
914 : : }
915 : :
916 : : /*
917 : : * fileIsForeignScanParallelSafe
918 : : * Reading a file, or external program, in a parallel worker should work
919 : : * just the same as reading it in the leader, so mark scans safe.
920 : : */
921 : : static bool
922 : 39 : fileIsForeignScanParallelSafe(PlannerInfo *root, RelOptInfo *rel,
923 : : RangeTblEntry *rte)
924 : : {
925 : 39 : return true;
926 : : }
927 : :
928 : : /*
929 : : * check_selective_binary_conversion
930 : : *
931 : : * Check to see if it's useful to convert only a subset of the file's columns
932 : : * to binary. If so, construct a list of the column names to be converted,
933 : : * return that at *columns, and return true. (Note that it's possible to
934 : : * determine that no columns need be converted, for instance with a COUNT(*)
935 : : * query. So we can't use returning a NIL list to indicate failure.)
936 : : */
937 : : static bool
938 : 43 : check_selective_binary_conversion(RelOptInfo *baserel,
939 : : Oid foreigntableid,
940 : : List **columns)
941 : : {
942 : : ForeignTable *table;
943 : : ListCell *lc;
944 : : Relation rel;
945 : : TupleDesc tupleDesc;
946 : : int attidx;
947 : 43 : Bitmapset *attrs_used = NULL;
948 : 43 : bool has_wholerow = false;
949 : : int numattrs;
950 : : int i;
951 : :
952 : 43 : *columns = NIL; /* default result */
953 : :
954 : : /*
955 : : * Check format of the file. If binary format, this is irrelevant.
956 : : */
957 : 43 : table = GetForeignTable(foreigntableid);
958 [ + - + - : 43 : foreach(lc, table->options)
+ - ]
959 : : {
960 : 43 : DefElem *def = (DefElem *) lfirst(lc);
961 : :
962 [ + - ]: 43 : if (strcmp(def->defname, "format") == 0)
963 : : {
964 : 43 : char *format = defGetString(def);
965 : :
966 [ - + ]: 43 : if (strcmp(format, "binary") == 0)
967 : 0 : return false;
968 : 43 : break;
969 : : }
970 : : }
971 : :
972 : : /* Collect all the attributes needed for joins or final output. */
973 : 43 : pull_varattnos((Node *) baserel->reltarget->exprs, baserel->relid,
974 : : &attrs_used);
975 : :
976 : : /* Add all the attributes used by restriction clauses. */
977 [ + + + + : 51 : foreach(lc, baserel->baserestrictinfo)
+ + ]
978 : : {
979 : 8 : RestrictInfo *rinfo = (RestrictInfo *) lfirst(lc);
980 : :
981 : 8 : pull_varattnos((Node *) rinfo->clause, baserel->relid,
982 : : &attrs_used);
983 : : }
984 : :
985 : : /* Convert attribute numbers to column names. */
986 : 43 : rel = table_open(foreigntableid, AccessShareLock);
987 : 43 : tupleDesc = RelationGetDescr(rel);
988 : :
989 : 43 : attidx = -1;
990 [ + + ]: 134 : while ((attidx = bms_next_member(attrs_used, attidx)) >= 0)
991 : : {
992 : : /* attidx is zero-based, attnum is the normal attribute number */
993 : 95 : AttrNumber attnum = attidx + FirstLowInvalidHeapAttributeNumber;
994 : :
995 [ + + ]: 95 : if (attnum == 0)
996 : : {
997 : 4 : has_wholerow = true;
998 : 4 : break;
999 : : }
1000 : :
1001 : : /* Ignore system attributes. */
1002 [ + + ]: 91 : if (attnum < 0)
1003 : 13 : continue;
1004 : :
1005 : : /* Get user attributes. */
1006 [ + - ]: 78 : if (attnum > 0)
1007 : : {
1008 : 78 : Form_pg_attribute attr = TupleDescAttr(tupleDesc, attnum - 1);
1009 : 78 : char *attname = NameStr(attr->attname);
1010 : :
1011 : : /* Skip dropped attributes (probably shouldn't see any here). */
1012 [ - + ]: 78 : if (attr->attisdropped)
1013 : 0 : continue;
1014 : :
1015 : : /*
1016 : : * Skip generated columns (COPY won't accept them in the column
1017 : : * list)
1018 : : */
1019 [ + + ]: 78 : if (attr->attgenerated)
1020 : 1 : continue;
1021 : 77 : *columns = lappend(*columns, makeString(pstrdup(attname)));
1022 : : }
1023 : : }
1024 : :
1025 : : /* Count non-dropped user attributes while we have the tupdesc. */
1026 : 43 : numattrs = 0;
1027 [ + + ]: 135 : for (i = 0; i < tupleDesc->natts; i++)
1028 : : {
1029 [ - + ]: 92 : if (TupleDescCompactAttr(tupleDesc, i)->attisdropped)
1030 : 0 : continue;
1031 : 92 : numattrs++;
1032 : : }
1033 : :
1034 : 43 : table_close(rel, AccessShareLock);
1035 : :
1036 : : /* If there's a whole-row reference, fail: we need all the columns. */
1037 [ + + ]: 43 : if (has_wholerow)
1038 : : {
1039 : 4 : *columns = NIL;
1040 : 4 : return false;
1041 : : }
1042 : :
1043 : : /* If all the user attributes are needed, fail. */
1044 [ + + ]: 39 : if (numattrs == list_length(*columns))
1045 : : {
1046 : 34 : *columns = NIL;
1047 : 34 : return false;
1048 : : }
1049 : :
1050 : 5 : return true;
1051 : : }
1052 : :
1053 : : /*
1054 : : * Estimate size of a foreign table.
1055 : : *
1056 : : * The main result is returned in baserel->rows. We also set
1057 : : * fdw_private->pages and fdw_private->ntuples for later use in the cost
1058 : : * calculation.
1059 : : */
1060 : : static void
1061 : 43 : estimate_size(PlannerInfo *root, RelOptInfo *baserel,
1062 : : FileFdwPlanState *fdw_private)
1063 : : {
1064 : : struct stat stat_buf;
1065 : : BlockNumber pages;
1066 : : double ntuples;
1067 : : double nrows;
1068 : :
1069 : : /*
1070 : : * Get size of the file. It might not be there at plan time, though, in
1071 : : * which case we have to use a default estimate. We also have to fall
1072 : : * back to the default if using a program as the input.
1073 : : */
1074 [ + - - + ]: 43 : if (fdw_private->is_program || stat(fdw_private->filename, &stat_buf) < 0)
1075 : 0 : stat_buf.st_size = 10 * BLCKSZ;
1076 : :
1077 : : /*
1078 : : * Convert size to pages for use in I/O cost estimate later.
1079 : : */
1080 : 43 : pages = (stat_buf.st_size + (BLCKSZ - 1)) / BLCKSZ;
1081 [ - + ]: 43 : if (pages < 1)
1082 : 0 : pages = 1;
1083 : 43 : fdw_private->pages = pages;
1084 : :
1085 : : /*
1086 : : * Estimate the number of tuples in the file.
1087 : : */
1088 [ - + - - ]: 43 : if (baserel->tuples >= 0 && baserel->pages > 0)
1089 : 0 : {
1090 : : /*
1091 : : * We have # of pages and # of tuples from pg_class (that is, from a
1092 : : * previous ANALYZE), so compute a tuples-per-page estimate and scale
1093 : : * that by the current file size.
1094 : : */
1095 : : double density;
1096 : :
1097 : 0 : density = baserel->tuples / (double) baserel->pages;
1098 : 0 : ntuples = clamp_row_est(density * (double) pages);
1099 : : }
1100 : : else
1101 : : {
1102 : : /*
1103 : : * Otherwise we have to fake it. We back into this estimate using the
1104 : : * planner's idea of the relation width; which is bogus if not all
1105 : : * columns are being read, not to mention that the text representation
1106 : : * of a row probably isn't the same size as its internal
1107 : : * representation. Possibly we could do something better, but the
1108 : : * real answer to anyone who complains is "ANALYZE" ...
1109 : : */
1110 : : int tuple_width;
1111 : :
1112 : 43 : tuple_width = MAXALIGN(baserel->reltarget->width) +
1113 : : MAXALIGN(SizeofHeapTupleHeader);
1114 : 43 : ntuples = clamp_row_est((double) stat_buf.st_size /
1115 : 43 : (double) tuple_width);
1116 : : }
1117 : 43 : fdw_private->ntuples = ntuples;
1118 : :
1119 : : /*
1120 : : * Now estimate the number of rows returned by the scan after applying the
1121 : : * baserestrictinfo quals.
1122 : : */
1123 : 43 : nrows = ntuples *
1124 : 43 : clauselist_selectivity(root,
1125 : : baserel->baserestrictinfo,
1126 : : 0,
1127 : : JOIN_INNER,
1128 : : NULL);
1129 : :
1130 : 43 : nrows = clamp_row_est(nrows);
1131 : :
1132 : : /* Save the output-rows estimate for the planner */
1133 : 43 : baserel->rows = nrows;
1134 : 43 : }
1135 : :
1136 : : /*
1137 : : * Estimate costs of scanning a foreign table.
1138 : : *
1139 : : * Results are returned in *startup_cost and *total_cost.
1140 : : */
1141 : : static void
1142 : 43 : estimate_costs(PlannerInfo *root, RelOptInfo *baserel,
1143 : : FileFdwPlanState *fdw_private,
1144 : : Cost *startup_cost, Cost *total_cost)
1145 : : {
1146 : 43 : BlockNumber pages = fdw_private->pages;
1147 : 43 : double ntuples = fdw_private->ntuples;
1148 : 43 : Cost run_cost = 0;
1149 : : Cost cpu_per_tuple;
1150 : :
1151 : : /*
1152 : : * We estimate costs almost the same way as cost_seqscan(), thus assuming
1153 : : * that I/O costs are equivalent to a regular table file of the same size.
1154 : : * However, we take per-tuple CPU costs as 10x of a seqscan, to account
1155 : : * for the cost of parsing records.
1156 : : *
1157 : : * In the case of a program source, this calculation is even more divorced
1158 : : * from reality, but we have no good alternative; and it's not clear that
1159 : : * the numbers we produce here matter much anyway, since there's only one
1160 : : * access path for the rel.
1161 : : */
1162 : 43 : run_cost += seq_page_cost * pages;
1163 : :
1164 : 43 : *startup_cost = baserel->baserestrictcost.startup;
1165 : 43 : cpu_per_tuple = cpu_tuple_cost * 10 + baserel->baserestrictcost.per_tuple;
1166 : 43 : run_cost += cpu_per_tuple * ntuples;
1167 : 43 : *total_cost = *startup_cost + run_cost;
1168 : 43 : }
1169 : :
1170 : : /*
1171 : : * file_acquire_sample_rows -- acquire a random sample of rows from the table
1172 : : *
1173 : : * Selected rows are returned in the caller-allocated array rows[],
1174 : : * which must have at least targrows entries.
1175 : : * The actual number of rows selected is returned as the function result.
1176 : : * We also count the total number of rows in the file and return it into
1177 : : * *totalrows. Rows skipped due to on_error = 'ignore' are not included
1178 : : * in this count. Note that *totaldeadrows is always set to 0.
1179 : : *
1180 : : * Note that the returned list of rows is not always in order by physical
1181 : : * position in the file. Therefore, correlation estimates derived later
1182 : : * may be meaningless, but it's OK because we don't use the estimates
1183 : : * currently (the planner only pays attention to correlation for indexscans).
1184 : : */
1185 : : static int
1186 : 1 : file_acquire_sample_rows(Relation onerel, int elevel,
1187 : : HeapTuple *rows, int targrows,
1188 : : double *totalrows, double *totaldeadrows)
1189 : : {
1190 : 1 : int numrows = 0;
1191 : 1 : double rowstoskip = -1; /* -1 means not set yet */
1192 : : ReservoirStateData rstate;
1193 : : TupleDesc tupDesc;
1194 : : Datum *values;
1195 : : bool *nulls;
1196 : : bool found;
1197 : : char *filename;
1198 : : bool is_program;
1199 : : List *options;
1200 : : CopyFromState cstate;
1201 : : ErrorContextCallback errcallback;
1202 : 1 : MemoryContext oldcontext = CurrentMemoryContext;
1203 : : MemoryContext tupcontext;
1204 : :
1205 : : Assert(onerel);
1206 : : Assert(targrows > 0);
1207 : :
1208 : 1 : tupDesc = RelationGetDescr(onerel);
1209 : 1 : values = (Datum *) palloc(tupDesc->natts * sizeof(Datum));
1210 : 1 : nulls = (bool *) palloc(tupDesc->natts * sizeof(bool));
1211 : :
1212 : : /* Fetch options of foreign table */
1213 : 1 : fileGetOptions(RelationGetRelid(onerel), &filename, &is_program, &options);
1214 : :
1215 : : /*
1216 : : * Create CopyState from FDW options.
1217 : : */
1218 : 1 : cstate = BeginCopyFrom(NULL, onerel, NULL, filename, is_program, NULL, NIL,
1219 : : options);
1220 : :
1221 : : /*
1222 : : * Use per-tuple memory context to prevent leak of memory used to read
1223 : : * rows from the file with Copy routines.
1224 : : */
1225 : 1 : tupcontext = AllocSetContextCreate(CurrentMemoryContext,
1226 : : "file_fdw temporary context",
1227 : : ALLOCSET_DEFAULT_SIZES);
1228 : :
1229 : : /* Prepare for sampling rows */
1230 : 1 : reservoir_init_selection_state(&rstate, targrows);
1231 : :
1232 : : /* Set up callback to identify error line number. */
1233 : 1 : errcallback.callback = CopyFromErrorCallback;
1234 : 1 : errcallback.arg = cstate;
1235 : 1 : errcallback.previous = error_context_stack;
1236 : 1 : error_context_stack = &errcallback;
1237 : :
1238 : 1 : *totalrows = 0;
1239 : 1 : *totaldeadrows = 0;
1240 : : for (;;)
1241 : : {
1242 : : /* Check for user-requested abort or sleep */
1243 : 5 : vacuum_delay_point(true);
1244 : :
1245 : : /* Fetch next row */
1246 : 5 : MemoryContextReset(tupcontext);
1247 : 5 : MemoryContextSwitchTo(tupcontext);
1248 : :
1249 : 5 : found = NextCopyFrom(cstate, NULL, values, nulls);
1250 : :
1251 : 5 : MemoryContextSwitchTo(oldcontext);
1252 : :
1253 [ + + ]: 5 : if (!found)
1254 : 1 : break;
1255 : :
1256 [ + - ]: 4 : if (cstate->opts.on_error == COPY_ON_ERROR_IGNORE &&
1257 [ + + ]: 4 : cstate->escontext->error_occurred)
1258 : : {
1259 : : /*
1260 : : * Soft error occurred, skip this tuple and just make
1261 : : * ErrorSaveContext ready for the next NextCopyFrom. Since we
1262 : : * don't set details_wanted and error_data is not to be filled,
1263 : : * just resetting error_occurred is enough.
1264 : : */
1265 : 2 : cstate->escontext->error_occurred = false;
1266 : :
1267 : : /* Repeat NextCopyFrom() until no soft error occurs */
1268 : 2 : continue;
1269 : : }
1270 : :
1271 : : /*
1272 : : * The first targrows sample rows are simply copied into the
1273 : : * reservoir. Then we start replacing tuples in the sample until we
1274 : : * reach the end of the relation. This algorithm is from Jeff Vitter's
1275 : : * paper (see more info in commands/analyze.c).
1276 : : */
1277 [ + - ]: 2 : if (numrows < targrows)
1278 : : {
1279 : 2 : rows[numrows++] = heap_form_tuple(tupDesc, values, nulls);
1280 : : }
1281 : : else
1282 : : {
1283 : : /*
1284 : : * t in Vitter's paper is the number of records already processed.
1285 : : * If we need to compute a new S value, we must use the
1286 : : * not-yet-incremented value of totalrows as t.
1287 : : */
1288 [ # # ]: 0 : if (rowstoskip < 0)
1289 : 0 : rowstoskip = reservoir_get_next_S(&rstate, *totalrows, targrows);
1290 : :
1291 [ # # ]: 0 : if (rowstoskip <= 0)
1292 : : {
1293 : : /*
1294 : : * Found a suitable tuple, so save it, replacing one old tuple
1295 : : * at random
1296 : : */
1297 : 0 : int k = (int) (targrows * sampler_random_fract(&rstate.randstate));
1298 : :
1299 : : Assert(k >= 0 && k < targrows);
1300 : 0 : heap_freetuple(rows[k]);
1301 : 0 : rows[k] = heap_form_tuple(tupDesc, values, nulls);
1302 : : }
1303 : :
1304 : 0 : rowstoskip -= 1;
1305 : : }
1306 : :
1307 : 2 : *totalrows += 1;
1308 : : }
1309 : :
1310 : : /* Remove error callback. */
1311 : 1 : error_context_stack = errcallback.previous;
1312 : :
1313 : : /* Clean up. */
1314 : 1 : MemoryContextDelete(tupcontext);
1315 : :
1316 [ + - ]: 1 : if (cstate->opts.on_error == COPY_ON_ERROR_IGNORE &&
1317 [ + - ]: 1 : cstate->num_errors > 0 &&
1318 [ - + ]: 1 : cstate->opts.log_verbosity >= COPY_LOG_VERBOSITY_DEFAULT)
1319 [ # # ]: 0 : ereport(NOTICE,
1320 : : errmsg_plural("%" PRIu64 " row was skipped due to data type incompatibility",
1321 : : "%" PRIu64 " rows were skipped due to data type incompatibility",
1322 : : cstate->num_errors,
1323 : : cstate->num_errors));
1324 : :
1325 : 1 : EndCopyFrom(cstate);
1326 : :
1327 : 1 : pfree(values);
1328 : 1 : pfree(nulls);
1329 : :
1330 : : /*
1331 : : * Emit some interesting relation info
1332 : : */
1333 [ - + ]: 1 : ereport(elevel,
1334 : : (errmsg("\"%s\": file contains %.0f rows; "
1335 : : "%d rows in sample",
1336 : : RelationGetRelationName(onerel),
1337 : : *totalrows, numrows)));
1338 : :
1339 : 1 : return numrows;
1340 : : }
|
