LCOV - code coverage report
Current view: top level - contrib/file_fdw - file_fdw.c (source / functions) Hit Total Coverage
Test: PostgreSQL 18devel Lines: 312 347 89.9 %
Date: 2024-12-12 17:14:55 Functions: 22 22 100.0 %
Legend: Lines: hit not hit

          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) 2010-2024, 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.h"
      28             : #include "commands/vacuum.h"
      29             : #include "foreign/fdwapi.h"
      30             : #include "foreign/foreign.h"
      31             : #include "miscadmin.h"
      32             : #include "nodes/makefuncs.h"
      33             : #include "optimizer/optimizer.h"
      34             : #include "optimizer/pathnode.h"
      35             : #include "optimizer/planmain.h"
      36             : #include "optimizer/restrictinfo.h"
      37             : #include "utils/acl.h"
      38             : #include "utils/memutils.h"
      39             : #include "utils/rel.h"
      40             : #include "utils/sampling.h"
      41             : #include "utils/varlena.h"
      42             : 
      43           2 : PG_MODULE_MAGIC;
      44             : 
      45             : /*
      46             :  * Describes the valid options for objects that use this wrapper.
      47             :  */
      48             : struct FileFdwOption
      49             : {
      50             :     const char *optname;
      51             :     Oid         optcontext;     /* Oid of catalog in which option may appear */
      52             : };
      53             : 
      54             : /*
      55             :  * Valid options for file_fdw.
      56             :  * These options are based on the options for the COPY FROM command.
      57             :  * But note that force_not_null and force_null are handled as boolean options
      58             :  * attached to a column, not as table options.
      59             :  *
      60             :  * Note: If you are adding new option for user mapping, you need to modify
      61             :  * fileGetOptions(), which currently doesn't bother to look at user mappings.
      62             :  */
      63             : static const struct FileFdwOption valid_options[] = {
      64             :     /* Data source options */
      65             :     {"filename", ForeignTableRelationId},
      66             :     {"program", ForeignTableRelationId},
      67             : 
      68             :     /* Format options */
      69             :     /* oids option is not supported */
      70             :     {"format", ForeignTableRelationId},
      71             :     {"header", ForeignTableRelationId},
      72             :     {"delimiter", ForeignTableRelationId},
      73             :     {"quote", ForeignTableRelationId},
      74             :     {"escape", ForeignTableRelationId},
      75             :     {"null", ForeignTableRelationId},
      76             :     {"default", ForeignTableRelationId},
      77             :     {"encoding", ForeignTableRelationId},
      78             :     {"on_error", ForeignTableRelationId},
      79             :     {"log_verbosity", ForeignTableRelationId},
      80             :     {"reject_limit", ForeignTableRelationId},
      81             :     {"force_not_null", AttributeRelationId},
      82             :     {"force_null", AttributeRelationId},
      83             : 
      84             :     /*
      85             :      * force_quote is not supported by file_fdw because it's for COPY TO.
      86             :      */
      87             : 
      88             :     /* Sentinel */
      89             :     {NULL, InvalidOid}
      90             : };
      91             : 
      92             : /*
      93             :  * FDW-specific information for RelOptInfo.fdw_private.
      94             :  */
      95             : typedef struct FileFdwPlanState
      96             : {
      97             :     char       *filename;       /* file or program to read from */
      98             :     bool        is_program;     /* true if filename represents an OS command */
      99             :     List       *options;        /* merged COPY options, excluding filename and
     100             :                                  * is_program */
     101             :     BlockNumber pages;          /* estimate of file's physical size */
     102             :     double      ntuples;        /* estimate of number of data rows */
     103             : } FileFdwPlanState;
     104             : 
     105             : /*
     106             :  * FDW-specific information for ForeignScanState.fdw_state.
     107             :  */
     108             : typedef struct FileFdwExecutionState
     109             : {
     110             :     char       *filename;       /* file or program to read from */
     111             :     bool        is_program;     /* true if filename represents an OS command */
     112             :     List       *options;        /* merged COPY options, excluding filename and
     113             :                                  * is_program */
     114             :     CopyFromState cstate;       /* COPY execution state */
     115             : } FileFdwExecutionState;
     116             : 
     117             : /*
     118             :  * SQL functions
     119             :  */
     120           4 : PG_FUNCTION_INFO_V1(file_fdw_handler);
     121           4 : PG_FUNCTION_INFO_V1(file_fdw_validator);
     122             : 
     123             : /*
     124             :  * FDW callback routines
     125             :  */
     126             : static void fileGetForeignRelSize(PlannerInfo *root,
     127             :                                   RelOptInfo *baserel,
     128             :                                   Oid foreigntableid);
     129             : static void fileGetForeignPaths(PlannerInfo *root,
     130             :                                 RelOptInfo *baserel,
     131             :                                 Oid foreigntableid);
     132             : static ForeignScan *fileGetForeignPlan(PlannerInfo *root,
     133             :                                        RelOptInfo *baserel,
     134             :                                        Oid foreigntableid,
     135             :                                        ForeignPath *best_path,
     136             :                                        List *tlist,
     137             :                                        List *scan_clauses,
     138             :                                        Plan *outer_plan);
     139             : static void fileExplainForeignScan(ForeignScanState *node, ExplainState *es);
     140             : static void fileBeginForeignScan(ForeignScanState *node, int eflags);
     141             : static TupleTableSlot *fileIterateForeignScan(ForeignScanState *node);
     142             : static void fileReScanForeignScan(ForeignScanState *node);
     143             : static void fileEndForeignScan(ForeignScanState *node);
     144             : static bool fileAnalyzeForeignTable(Relation relation,
     145             :                                     AcquireSampleRowsFunc *func,
     146             :                                     BlockNumber *totalpages);
     147             : static bool fileIsForeignScanParallelSafe(PlannerInfo *root, RelOptInfo *rel,
     148             :                                           RangeTblEntry *rte);
     149             : 
     150             : /*
     151             :  * Helper functions
     152             :  */
     153             : static bool is_valid_option(const char *option, Oid context);
     154             : static void fileGetOptions(Oid foreigntableid,
     155             :                            char **filename,
     156             :                            bool *is_program,
     157             :                            List **other_options);
     158             : static List *get_file_fdw_attribute_options(Oid relid);
     159             : static bool check_selective_binary_conversion(RelOptInfo *baserel,
     160             :                                               Oid foreigntableid,
     161             :                                               List **columns);
     162             : static void estimate_size(PlannerInfo *root, RelOptInfo *baserel,
     163             :                           FileFdwPlanState *fdw_private);
     164             : static void estimate_costs(PlannerInfo *root, RelOptInfo *baserel,
     165             :                            FileFdwPlanState *fdw_private,
     166             :                            Cost *startup_cost, Cost *total_cost);
     167             : static int  file_acquire_sample_rows(Relation onerel, int elevel,
     168             :                                      HeapTuple *rows, int targrows,
     169             :                                      double *totalrows, double *totaldeadrows);
     170             : 
     171             : 
     172             : /*
     173             :  * Foreign-data wrapper handler function: return a struct with pointers
     174             :  * to my callback routines.
     175             :  */
     176             : Datum
     177          38 : file_fdw_handler(PG_FUNCTION_ARGS)
     178             : {
     179          38 :     FdwRoutine *fdwroutine = makeNode(FdwRoutine);
     180             : 
     181          38 :     fdwroutine->GetForeignRelSize = fileGetForeignRelSize;
     182          38 :     fdwroutine->GetForeignPaths = fileGetForeignPaths;
     183          38 :     fdwroutine->GetForeignPlan = fileGetForeignPlan;
     184          38 :     fdwroutine->ExplainForeignScan = fileExplainForeignScan;
     185          38 :     fdwroutine->BeginForeignScan = fileBeginForeignScan;
     186          38 :     fdwroutine->IterateForeignScan = fileIterateForeignScan;
     187          38 :     fdwroutine->ReScanForeignScan = fileReScanForeignScan;
     188          38 :     fdwroutine->EndForeignScan = fileEndForeignScan;
     189          38 :     fdwroutine->AnalyzeForeignTable = fileAnalyzeForeignTable;
     190          38 :     fdwroutine->IsForeignScanParallelSafe = fileIsForeignScanParallelSafe;
     191             : 
     192          38 :     PG_RETURN_POINTER(fdwroutine);
     193             : }
     194             : 
     195             : /*
     196             :  * Validate the generic options given to a FOREIGN DATA WRAPPER, SERVER,
     197             :  * USER MAPPING or FOREIGN TABLE that uses file_fdw.
     198             :  *
     199             :  * Raise an ERROR if the option or its value is considered invalid.
     200             :  */
     201             : Datum
     202         122 : file_fdw_validator(PG_FUNCTION_ARGS)
     203             : {
     204         122 :     List       *options_list = untransformRelOptions(PG_GETARG_DATUM(0));
     205         122 :     Oid         catalog = PG_GETARG_OID(1);
     206         122 :     char       *filename = NULL;
     207         122 :     DefElem    *force_not_null = NULL;
     208         122 :     DefElem    *force_null = NULL;
     209         122 :     List       *other_options = NIL;
     210             :     ListCell   *cell;
     211             : 
     212             :     /*
     213             :      * Check that only options supported by file_fdw, and allowed for the
     214             :      * current object type, are given.
     215             :      */
     216         402 :     foreach(cell, options_list)
     217             :     {
     218         298 :         DefElem    *def = (DefElem *) lfirst(cell);
     219             : 
     220         298 :         if (!is_valid_option(def->defname, catalog))
     221             :         {
     222             :             const struct FileFdwOption *opt;
     223             :             const char *closest_match;
     224             :             ClosestMatchState match_state;
     225          16 :             bool        has_valid_options = false;
     226             : 
     227             :             /*
     228             :              * Unknown option specified, complain about it. Provide a hint
     229             :              * with a valid option that looks similar, if there is one.
     230             :              */
     231          16 :             initClosestMatch(&match_state, def->defname, 4);
     232         256 :             for (opt = valid_options; opt->optname; opt++)
     233             :             {
     234         240 :                 if (catalog == opt->optcontext)
     235             :                 {
     236          52 :                     has_valid_options = true;
     237          52 :                     updateClosestMatch(&match_state, opt->optname);
     238             :                 }
     239             :             }
     240             : 
     241          16 :             closest_match = getClosestMatch(&match_state);
     242          16 :             ereport(ERROR,
     243             :                     (errcode(ERRCODE_FDW_INVALID_OPTION_NAME),
     244             :                      errmsg("invalid option \"%s\"", def->defname),
     245             :                      has_valid_options ? closest_match ?
     246             :                      errhint("Perhaps you meant the option \"%s\".",
     247             :                              closest_match) : 0 :
     248             :                      errhint("There are no valid options in this context.")));
     249             :         }
     250             : 
     251             :         /*
     252             :          * Separate out filename, program, and column-specific options, since
     253             :          * ProcessCopyOptions won't accept them.
     254             :          */
     255         282 :         if (strcmp(def->defname, "filename") == 0 ||
     256         250 :             strcmp(def->defname, "program") == 0)
     257             :         {
     258          32 :             if (filename)
     259           0 :                 ereport(ERROR,
     260             :                         (errcode(ERRCODE_SYNTAX_ERROR),
     261             :                          errmsg("conflicting or redundant options")));
     262             : 
     263             :             /*
     264             :              * Check permissions for changing which file or program is used by
     265             :              * the file_fdw.
     266             :              *
     267             :              * Only members of the role 'pg_read_server_files' are allowed to
     268             :              * set the 'filename' option of a file_fdw foreign table, while
     269             :              * only members of the role 'pg_execute_server_program' are
     270             :              * allowed to set the 'program' option.  This is because we don't
     271             :              * want regular users to be able to control which file gets read
     272             :              * or which program gets executed.
     273             :              *
     274             :              * Putting this sort of permissions check in a validator is a bit
     275             :              * of a crock, but there doesn't seem to be any other place that
     276             :              * can enforce the check more cleanly.
     277             :              *
     278             :              * Note that the valid_options[] array disallows setting filename
     279             :              * and program at any options level other than foreign table ---
     280             :              * otherwise there'd still be a security hole.
     281             :              */
     282          32 :             if (strcmp(def->defname, "filename") == 0 &&
     283          32 :                 !has_privs_of_role(GetUserId(), ROLE_PG_READ_SERVER_FILES))
     284           2 :                 ereport(ERROR,
     285             :                         (errcode(ERRCODE_INSUFFICIENT_PRIVILEGE),
     286             :                          errmsg("permission denied to set the \"%s\" option of a file_fdw foreign table",
     287             :                                 "filename"),
     288             :                          errdetail("Only roles with privileges of the \"%s\" role may set this option.",
     289             :                                    "pg_read_server_files")));
     290             : 
     291          30 :             if (strcmp(def->defname, "program") == 0 &&
     292           0 :                 !has_privs_of_role(GetUserId(), ROLE_PG_EXECUTE_SERVER_PROGRAM))
     293           0 :                 ereport(ERROR,
     294             :                         (errcode(ERRCODE_INSUFFICIENT_PRIVILEGE),
     295             :                          errmsg("permission denied to set the \"%s\" option of a file_fdw foreign table",
     296             :                                 "program"),
     297             :                          errdetail("Only roles with privileges of the \"%s\" role may set this option.",
     298             :                                    "pg_execute_server_program")));
     299             : 
     300          30 :             filename = defGetString(def);
     301             :         }
     302             : 
     303             :         /*
     304             :          * force_not_null is a boolean option; after validation we can discard
     305             :          * it - it will be retrieved later in get_file_fdw_attribute_options()
     306             :          */
     307         250 :         else if (strcmp(def->defname, "force_not_null") == 0)
     308             :         {
     309           8 :             if (force_not_null)
     310           0 :                 ereport(ERROR,
     311             :                         (errcode(ERRCODE_SYNTAX_ERROR),
     312             :                          errmsg("conflicting or redundant options"),
     313             :                          errhint("Option \"force_not_null\" supplied more than once for a column.")));
     314           8 :             force_not_null = def;
     315             :             /* Don't care what the value is, as long as it's a legal boolean */
     316           8 :             (void) defGetBoolean(def);
     317             :         }
     318             :         /* See comments for force_not_null above */
     319         242 :         else if (strcmp(def->defname, "force_null") == 0)
     320             :         {
     321           8 :             if (force_null)
     322           0 :                 ereport(ERROR,
     323             :                         (errcode(ERRCODE_SYNTAX_ERROR),
     324             :                          errmsg("conflicting or redundant options"),
     325             :                          errhint("Option \"force_null\" supplied more than once for a column.")));
     326           8 :             force_null = def;
     327           8 :             (void) defGetBoolean(def);
     328             :         }
     329             :         else
     330         234 :             other_options = lappend(other_options, def);
     331             :     }
     332             : 
     333             :     /*
     334             :      * Now apply the core COPY code's validation logic for more checks.
     335             :      */
     336         104 :     ProcessCopyOptions(NULL, NULL, true, other_options);
     337             : 
     338             :     /*
     339             :      * Either filename or program option is required for file_fdw foreign
     340             :      * tables.
     341             :      */
     342          54 :     if (catalog == ForeignTableRelationId && filename == NULL)
     343           2 :         ereport(ERROR,
     344             :                 (errcode(ERRCODE_FDW_DYNAMIC_PARAMETER_VALUE_NEEDED),
     345             :                  errmsg("either filename or program is required for file_fdw foreign tables")));
     346             : 
     347          52 :     PG_RETURN_VOID();
     348             : }
     349             : 
     350             : /*
     351             :  * Check if the provided option is one of the valid options.
     352             :  * context is the Oid of the catalog holding the object the option is for.
     353             :  */
     354             : static bool
     355         298 : is_valid_option(const char *option, Oid context)
     356             : {
     357             :     const struct FileFdwOption *opt;
     358             : 
     359        1876 :     for (opt = valid_options; opt->optname; opt++)
     360             :     {
     361        1860 :         if (context == opt->optcontext && strcmp(opt->optname, option) == 0)
     362         282 :             return true;
     363             :     }
     364          16 :     return false;
     365             : }
     366             : 
     367             : /*
     368             :  * Fetch the options for a file_fdw foreign table.
     369             :  *
     370             :  * We have to separate out filename/program from the other options because
     371             :  * those must not appear in the options list passed to the core COPY code.
     372             :  */
     373             : static void
     374         154 : fileGetOptions(Oid foreigntableid,
     375             :                char **filename, bool *is_program, List **other_options)
     376             : {
     377             :     ForeignTable *table;
     378             :     ForeignServer *server;
     379             :     ForeignDataWrapper *wrapper;
     380             :     List       *options;
     381             :     ListCell   *lc;
     382             : 
     383             :     /*
     384             :      * Extract options from FDW objects.  We ignore user mappings because
     385             :      * file_fdw doesn't have any options that can be specified there.
     386             :      *
     387             :      * (XXX Actually, given the current contents of valid_options[], there's
     388             :      * no point in examining anything except the foreign table's own options.
     389             :      * Simplify?)
     390             :      */
     391         154 :     table = GetForeignTable(foreigntableid);
     392         154 :     server = GetForeignServer(table->serverid);
     393         154 :     wrapper = GetForeignDataWrapper(server->fdwid);
     394             : 
     395         154 :     options = NIL;
     396         154 :     options = list_concat(options, wrapper->options);
     397         154 :     options = list_concat(options, server->options);
     398         154 :     options = list_concat(options, table->options);
     399         154 :     options = list_concat(options, get_file_fdw_attribute_options(foreigntableid));
     400             : 
     401             :     /*
     402             :      * Separate out the filename or program option (we assume there is only
     403             :      * one).
     404             :      */
     405         154 :     *filename = NULL;
     406         154 :     *is_program = false;
     407         308 :     foreach(lc, options)
     408             :     {
     409         308 :         DefElem    *def = (DefElem *) lfirst(lc);
     410             : 
     411         308 :         if (strcmp(def->defname, "filename") == 0)
     412             :         {
     413         154 :             *filename = defGetString(def);
     414         154 :             options = foreach_delete_current(options, lc);
     415         154 :             break;
     416             :         }
     417         154 :         else if (strcmp(def->defname, "program") == 0)
     418             :         {
     419           0 :             *filename = defGetString(def);
     420           0 :             *is_program = true;
     421           0 :             options = foreach_delete_current(options, lc);
     422           0 :             break;
     423             :         }
     424             :     }
     425             : 
     426             :     /*
     427             :      * The validator should have checked that filename or program was included
     428             :      * in the options, but check again, just in case.
     429             :      */
     430         154 :     if (*filename == NULL)
     431           0 :         elog(ERROR, "either filename or program is required for file_fdw foreign tables");
     432             : 
     433         154 :     *other_options = options;
     434         154 : }
     435             : 
     436             : /*
     437             :  * Retrieve per-column generic options from pg_attribute and construct a list
     438             :  * of DefElems representing them.
     439             :  *
     440             :  * At the moment we only have "force_not_null", and "force_null",
     441             :  * which should each be combined into a single DefElem listing all such
     442             :  * columns, since that's what COPY expects.
     443             :  */
     444             : static List *
     445         154 : get_file_fdw_attribute_options(Oid relid)
     446             : {
     447             :     Relation    rel;
     448             :     TupleDesc   tupleDesc;
     449             :     AttrNumber  natts;
     450             :     AttrNumber  attnum;
     451         154 :     List       *fnncolumns = NIL;
     452         154 :     List       *fncolumns = NIL;
     453             : 
     454         154 :     List       *options = NIL;
     455             : 
     456         154 :     rel = table_open(relid, AccessShareLock);
     457         154 :     tupleDesc = RelationGetDescr(rel);
     458         154 :     natts = tupleDesc->natts;
     459             : 
     460             :     /* Retrieve FDW options for all user-defined attributes. */
     461         486 :     for (attnum = 1; attnum <= natts; attnum++)
     462             :     {
     463         332 :         Form_pg_attribute attr = TupleDescAttr(tupleDesc, attnum - 1);
     464             :         List       *column_options;
     465             :         ListCell   *lc;
     466             : 
     467             :         /* Skip dropped attributes. */
     468         332 :         if (attr->attisdropped)
     469           0 :             continue;
     470             : 
     471         332 :         column_options = GetForeignColumnOptions(relid, attnum);
     472         364 :         foreach(lc, column_options)
     473             :         {
     474          32 :             DefElem    *def = (DefElem *) lfirst(lc);
     475             : 
     476          32 :             if (strcmp(def->defname, "force_not_null") == 0)
     477             :             {
     478          16 :                 if (defGetBoolean(def))
     479             :                 {
     480           8 :                     char       *attname = pstrdup(NameStr(attr->attname));
     481             : 
     482           8 :                     fnncolumns = lappend(fnncolumns, makeString(attname));
     483             :                 }
     484             :             }
     485          16 :             else if (strcmp(def->defname, "force_null") == 0)
     486             :             {
     487          16 :                 if (defGetBoolean(def))
     488             :                 {
     489           8 :                     char       *attname = pstrdup(NameStr(attr->attname));
     490             : 
     491           8 :                     fncolumns = lappend(fncolumns, makeString(attname));
     492             :                 }
     493             :             }
     494             :             /* maybe in future handle other column options here */
     495             :         }
     496             :     }
     497             : 
     498         154 :     table_close(rel, AccessShareLock);
     499             : 
     500             :     /*
     501             :      * Return DefElem only when some column(s) have force_not_null /
     502             :      * force_null options set
     503             :      */
     504         154 :     if (fnncolumns != NIL)
     505           8 :         options = lappend(options, makeDefElem("force_not_null", (Node *) fnncolumns, -1));
     506             : 
     507         154 :     if (fncolumns != NIL)
     508           8 :         options = lappend(options, makeDefElem("force_null", (Node *) fncolumns, -1));
     509             : 
     510         154 :     return options;
     511             : }
     512             : 
     513             : /*
     514             :  * fileGetForeignRelSize
     515             :  *      Obtain relation size estimates for a foreign table
     516             :  */
     517             : static void
     518          80 : fileGetForeignRelSize(PlannerInfo *root,
     519             :                       RelOptInfo *baserel,
     520             :                       Oid foreigntableid)
     521             : {
     522             :     FileFdwPlanState *fdw_private;
     523             : 
     524             :     /*
     525             :      * Fetch options.  We only need filename (or program) at this point, but
     526             :      * we might as well get everything and not need to re-fetch it later in
     527             :      * planning.
     528             :      */
     529          80 :     fdw_private = (FileFdwPlanState *) palloc(sizeof(FileFdwPlanState));
     530          80 :     fileGetOptions(foreigntableid,
     531             :                    &fdw_private->filename,
     532             :                    &fdw_private->is_program,
     533             :                    &fdw_private->options);
     534          80 :     baserel->fdw_private = fdw_private;
     535             : 
     536             :     /* Estimate relation size */
     537          80 :     estimate_size(root, baserel, fdw_private);
     538          80 : }
     539             : 
     540             : /*
     541             :  * fileGetForeignPaths
     542             :  *      Create possible access paths for a scan on the foreign table
     543             :  *
     544             :  *      Currently we don't support any push-down feature, so there is only one
     545             :  *      possible access path, which simply returns all records in the order in
     546             :  *      the data file.
     547             :  */
     548             : static void
     549          80 : fileGetForeignPaths(PlannerInfo *root,
     550             :                     RelOptInfo *baserel,
     551             :                     Oid foreigntableid)
     552             : {
     553          80 :     FileFdwPlanState *fdw_private = (FileFdwPlanState *) baserel->fdw_private;
     554             :     Cost        startup_cost;
     555             :     Cost        total_cost;
     556             :     List       *columns;
     557          80 :     List       *coptions = NIL;
     558             : 
     559             :     /* Decide whether to selectively perform binary conversion */
     560          80 :     if (check_selective_binary_conversion(baserel,
     561             :                                           foreigntableid,
     562             :                                           &columns))
     563           8 :         coptions = list_make1(makeDefElem("convert_selectively",
     564             :                                           (Node *) columns, -1));
     565             : 
     566             :     /* Estimate costs */
     567          80 :     estimate_costs(root, baserel, fdw_private,
     568             :                    &startup_cost, &total_cost);
     569             : 
     570             :     /*
     571             :      * Create a ForeignPath node and add it as only possible path.  We use the
     572             :      * fdw_private list of the path to carry the convert_selectively option;
     573             :      * it will be propagated into the fdw_private list of the Plan node.
     574             :      *
     575             :      * We don't support pushing join clauses into the quals of this path, but
     576             :      * it could still have required parameterization due to LATERAL refs in
     577             :      * its tlist.
     578             :      */
     579          80 :     add_path(baserel, (Path *)
     580          80 :              create_foreignscan_path(root, baserel,
     581             :                                      NULL,  /* default pathtarget */
     582             :                                      baserel->rows,
     583             :                                      0,
     584             :                                      startup_cost,
     585             :                                      total_cost,
     586             :                                      NIL,   /* no pathkeys */
     587             :                                      baserel->lateral_relids,
     588             :                                      NULL,  /* no extra plan */
     589             :                                      NIL,   /* no fdw_restrictinfo list */
     590             :                                      coptions));
     591             : 
     592             :     /*
     593             :      * If data file was sorted, and we knew it somehow, we could insert
     594             :      * appropriate pathkeys into the ForeignPath node to tell the planner
     595             :      * that.
     596             :      */
     597          80 : }
     598             : 
     599             : /*
     600             :  * fileGetForeignPlan
     601             :  *      Create a ForeignScan plan node for scanning the foreign table
     602             :  */
     603             : static ForeignScan *
     604          80 : fileGetForeignPlan(PlannerInfo *root,
     605             :                    RelOptInfo *baserel,
     606             :                    Oid foreigntableid,
     607             :                    ForeignPath *best_path,
     608             :                    List *tlist,
     609             :                    List *scan_clauses,
     610             :                    Plan *outer_plan)
     611             : {
     612          80 :     Index       scan_relid = baserel->relid;
     613             : 
     614             :     /*
     615             :      * We have no native ability to evaluate restriction clauses, so we just
     616             :      * put all the scan_clauses into the plan node's qual list for the
     617             :      * executor to check.  So all we have to do here is strip RestrictInfo
     618             :      * nodes from the clauses and ignore pseudoconstants (which will be
     619             :      * handled elsewhere).
     620             :      */
     621          80 :     scan_clauses = extract_actual_clauses(scan_clauses, false);
     622             : 
     623             :     /* Create the ForeignScan node */
     624          80 :     return make_foreignscan(tlist,
     625             :                             scan_clauses,
     626             :                             scan_relid,
     627             :                             NIL,    /* no expressions to evaluate */
     628             :                             best_path->fdw_private,
     629             :                             NIL,    /* no custom tlist */
     630             :                             NIL,    /* no remote quals */
     631             :                             outer_plan);
     632             : }
     633             : 
     634             : /*
     635             :  * fileExplainForeignScan
     636             :  *      Produce extra output for EXPLAIN
     637             :  */
     638             : static void
     639           6 : fileExplainForeignScan(ForeignScanState *node, ExplainState *es)
     640             : {
     641             :     char       *filename;
     642             :     bool        is_program;
     643             :     List       *options;
     644             : 
     645             :     /* Fetch options --- we only need filename and is_program at this point */
     646           6 :     fileGetOptions(RelationGetRelid(node->ss.ss_currentRelation),
     647             :                    &filename, &is_program, &options);
     648             : 
     649           6 :     if (is_program)
     650           0 :         ExplainPropertyText("Foreign Program", filename, es);
     651             :     else
     652           6 :         ExplainPropertyText("Foreign File", filename, es);
     653             : 
     654             :     /* Suppress file size if we're not showing cost details */
     655           6 :     if (es->costs)
     656             :     {
     657             :         struct stat stat_buf;
     658             : 
     659           0 :         if (!is_program &&
     660           0 :             stat(filename, &stat_buf) == 0)
     661           0 :             ExplainPropertyInteger("Foreign File Size", "b",
     662           0 :                                    (int64) stat_buf.st_size, es);
     663             :     }
     664           6 : }
     665             : 
     666             : /*
     667             :  * fileBeginForeignScan
     668             :  *      Initiate access to the file by creating CopyState
     669             :  */
     670             : static void
     671          70 : fileBeginForeignScan(ForeignScanState *node, int eflags)
     672             : {
     673          70 :     ForeignScan *plan = (ForeignScan *) node->ss.ps.plan;
     674             :     char       *filename;
     675             :     bool        is_program;
     676             :     List       *options;
     677             :     CopyFromState cstate;
     678             :     FileFdwExecutionState *festate;
     679             : 
     680             :     /*
     681             :      * Do nothing in EXPLAIN (no ANALYZE) case.  node->fdw_state stays NULL.
     682             :      */
     683          70 :     if (eflags & EXEC_FLAG_EXPLAIN_ONLY)
     684           6 :         return;
     685             : 
     686             :     /* Fetch options of foreign table */
     687          64 :     fileGetOptions(RelationGetRelid(node->ss.ss_currentRelation),
     688             :                    &filename, &is_program, &options);
     689             : 
     690             :     /* Add any options from the plan (currently only convert_selectively) */
     691          64 :     options = list_concat(options, plan->fdw_private);
     692             : 
     693             :     /*
     694             :      * Create CopyState from FDW options.  We always acquire all columns, so
     695             :      * as to match the expected ScanTupleSlot signature.
     696             :      */
     697          64 :     cstate = BeginCopyFrom(NULL,
     698             :                            node->ss.ss_currentRelation,
     699             :                            NULL,
     700             :                            filename,
     701             :                            is_program,
     702             :                            NULL,
     703             :                            NIL,
     704             :                            options);
     705             : 
     706             :     /*
     707             :      * Save state in node->fdw_state.  We must save enough information to call
     708             :      * BeginCopyFrom() again.
     709             :      */
     710          62 :     festate = (FileFdwExecutionState *) palloc(sizeof(FileFdwExecutionState));
     711          62 :     festate->filename = filename;
     712          62 :     festate->is_program = is_program;
     713          62 :     festate->options = options;
     714          62 :     festate->cstate = cstate;
     715             : 
     716          62 :     node->fdw_state = festate;
     717             : }
     718             : 
     719             : /*
     720             :  * fileIterateForeignScan
     721             :  *      Read next record from the data file and store it into the
     722             :  *      ScanTupleSlot as a virtual tuple
     723             :  */
     724             : static TupleTableSlot *
     725         246 : fileIterateForeignScan(ForeignScanState *node)
     726             : {
     727         246 :     FileFdwExecutionState *festate = (FileFdwExecutionState *) node->fdw_state;
     728         246 :     EState     *estate = CreateExecutorState();
     729             :     ExprContext *econtext;
     730         246 :     MemoryContext oldcontext = CurrentMemoryContext;
     731         246 :     TupleTableSlot *slot = node->ss.ss_ScanTupleSlot;
     732         246 :     CopyFromState cstate = festate->cstate;
     733             :     ErrorContextCallback errcallback;
     734             : 
     735             :     /* Set up callback to identify error line number. */
     736         246 :     errcallback.callback = CopyFromErrorCallback;
     737         246 :     errcallback.arg = cstate;
     738         246 :     errcallback.previous = error_context_stack;
     739         246 :     error_context_stack = &errcallback;
     740             : 
     741             :     /*
     742             :      * We pass ExprContext because there might be a use of the DEFAULT option
     743             :      * in COPY FROM, so we may need to evaluate default expressions.
     744             :      */
     745         246 :     econtext = GetPerTupleExprContext(estate);
     746             : 
     747         260 : retry:
     748             : 
     749             :     /*
     750             :      * DEFAULT expressions need to be evaluated in a per-tuple context, so
     751             :      * switch in case we are doing that.
     752             :      */
     753         260 :     MemoryContextSwitchTo(GetPerTupleMemoryContext(estate));
     754             : 
     755             :     /*
     756             :      * The protocol for loading a virtual tuple into a slot is first
     757             :      * ExecClearTuple, then fill the values/isnull arrays, then
     758             :      * ExecStoreVirtualTuple.  If we don't find another row in the file, we
     759             :      * just skip the last step, leaving the slot empty as required.
     760             :      *
     761             :      */
     762         260 :     ExecClearTuple(slot);
     763             : 
     764         260 :     if (NextCopyFrom(cstate, econtext, slot->tts_values, slot->tts_isnull))
     765             :     {
     766         196 :         if (cstate->opts.on_error == COPY_ON_ERROR_IGNORE &&
     767          32 :             cstate->escontext->error_occurred)
     768             :         {
     769             :             /*
     770             :              * Soft error occurred, skip this tuple and just make
     771             :              * ErrorSaveContext ready for the next NextCopyFrom. Since we
     772             :              * don't set details_wanted and error_data is not to be filled,
     773             :              * just resetting error_occurred is enough.
     774             :              */
     775          16 :             cstate->escontext->error_occurred = false;
     776             : 
     777             :             /* Switch back to original memory context */
     778          16 :             MemoryContextSwitchTo(oldcontext);
     779             : 
     780             :             /*
     781             :              * Make sure we are interruptible while repeatedly calling
     782             :              * NextCopyFrom() until no soft error occurs.
     783             :              */
     784          16 :             CHECK_FOR_INTERRUPTS();
     785             : 
     786             :             /*
     787             :              * Reset the per-tuple exprcontext, to clean-up after expression
     788             :              * evaluations etc.
     789             :              */
     790          16 :             ResetPerTupleExprContext(estate);
     791             : 
     792          16 :             if (cstate->opts.reject_limit > 0 &&
     793           8 :                 cstate->num_errors > cstate->opts.reject_limit)
     794           2 :                 ereport(ERROR,
     795             :                         (errcode(ERRCODE_INVALID_TEXT_REPRESENTATION),
     796             :                          errmsg("skipped more than REJECT_LIMIT (%lld) rows due to data type incompatibility",
     797             :                                 (long long) cstate->opts.reject_limit)));
     798             : 
     799             :             /* Repeat NextCopyFrom() until no soft error occurs */
     800          14 :             goto retry;
     801             :         }
     802             : 
     803         180 :         ExecStoreVirtualTuple(slot);
     804             :     }
     805             : 
     806             :     /* Switch back to original memory context */
     807         240 :     MemoryContextSwitchTo(oldcontext);
     808             : 
     809             :     /* Remove error callback. */
     810         240 :     error_context_stack = errcallback.previous;
     811             : 
     812         240 :     return slot;
     813             : }
     814             : 
     815             : /*
     816             :  * fileReScanForeignScan
     817             :  *      Rescan table, possibly with new parameters
     818             :  */
     819             : static void
     820           6 : fileReScanForeignScan(ForeignScanState *node)
     821             : {
     822           6 :     FileFdwExecutionState *festate = (FileFdwExecutionState *) node->fdw_state;
     823             : 
     824           6 :     EndCopyFrom(festate->cstate);
     825             : 
     826          12 :     festate->cstate = BeginCopyFrom(NULL,
     827             :                                     node->ss.ss_currentRelation,
     828             :                                     NULL,
     829           6 :                                     festate->filename,
     830           6 :                                     festate->is_program,
     831             :                                     NULL,
     832             :                                     NIL,
     833             :                                     festate->options);
     834           6 : }
     835             : 
     836             : /*
     837             :  * fileEndForeignScan
     838             :  *      Finish scanning foreign table and dispose objects used for this scan
     839             :  */
     840             : static void
     841          62 : fileEndForeignScan(ForeignScanState *node)
     842             : {
     843          62 :     FileFdwExecutionState *festate = (FileFdwExecutionState *) node->fdw_state;
     844             : 
     845             :     /* if festate is NULL, we are in EXPLAIN; nothing to do */
     846          62 :     if (!festate)
     847           6 :         return;
     848             : 
     849          56 :     if (festate->cstate->opts.on_error == COPY_ON_ERROR_IGNORE &&
     850           6 :         festate->cstate->num_errors > 0 &&
     851           6 :         festate->cstate->opts.log_verbosity >= COPY_LOG_VERBOSITY_DEFAULT)
     852           2 :         ereport(NOTICE,
     853             :                 errmsg_plural("%llu row was skipped due to data type incompatibility",
     854             :                               "%llu rows were skipped due to data type incompatibility",
     855             :                               (unsigned long long) festate->cstate->num_errors,
     856             :                               (unsigned long long) festate->cstate->num_errors));
     857             : 
     858          56 :     EndCopyFrom(festate->cstate);
     859             : }
     860             : 
     861             : /*
     862             :  * fileAnalyzeForeignTable
     863             :  *      Test whether analyzing this foreign table is supported
     864             :  */
     865             : static bool
     866           2 : fileAnalyzeForeignTable(Relation relation,
     867             :                         AcquireSampleRowsFunc *func,
     868             :                         BlockNumber *totalpages)
     869             : {
     870             :     char       *filename;
     871             :     bool        is_program;
     872             :     List       *options;
     873             :     struct stat stat_buf;
     874             : 
     875             :     /* Fetch options of foreign table */
     876           2 :     fileGetOptions(RelationGetRelid(relation), &filename, &is_program, &options);
     877             : 
     878             :     /*
     879             :      * If this is a program instead of a file, just return false to skip
     880             :      * analyzing the table.  We could run the program and collect stats on
     881             :      * whatever it currently returns, but it seems likely that in such cases
     882             :      * the output would be too volatile for the stats to be useful.  Maybe
     883             :      * there should be an option to enable doing this?
     884             :      */
     885           2 :     if (is_program)
     886           0 :         return false;
     887             : 
     888             :     /*
     889             :      * Get size of the file.  (XXX if we fail here, would it be better to just
     890             :      * return false to skip analyzing the table?)
     891             :      */
     892           2 :     if (stat(filename, &stat_buf) < 0)
     893           0 :         ereport(ERROR,
     894             :                 (errcode_for_file_access(),
     895             :                  errmsg("could not stat file \"%s\": %m",
     896             :                         filename)));
     897             : 
     898             :     /*
     899             :      * Convert size to pages.  Must return at least 1 so that we can tell
     900             :      * later on that pg_class.relpages is not default.
     901             :      */
     902           2 :     *totalpages = (stat_buf.st_size + (BLCKSZ - 1)) / BLCKSZ;
     903           2 :     if (*totalpages < 1)
     904           0 :         *totalpages = 1;
     905             : 
     906           2 :     *func = file_acquire_sample_rows;
     907             : 
     908           2 :     return true;
     909             : }
     910             : 
     911             : /*
     912             :  * fileIsForeignScanParallelSafe
     913             :  *      Reading a file, or external program, in a parallel worker should work
     914             :  *      just the same as reading it in the leader, so mark scans safe.
     915             :  */
     916             : static bool
     917          72 : fileIsForeignScanParallelSafe(PlannerInfo *root, RelOptInfo *rel,
     918             :                               RangeTblEntry *rte)
     919             : {
     920          72 :     return true;
     921             : }
     922             : 
     923             : /*
     924             :  * check_selective_binary_conversion
     925             :  *
     926             :  * Check to see if it's useful to convert only a subset of the file's columns
     927             :  * to binary.  If so, construct a list of the column names to be converted,
     928             :  * return that at *columns, and return true.  (Note that it's possible to
     929             :  * determine that no columns need be converted, for instance with a COUNT(*)
     930             :  * query.  So we can't use returning a NIL list to indicate failure.)
     931             :  */
     932             : static bool
     933          80 : check_selective_binary_conversion(RelOptInfo *baserel,
     934             :                                   Oid foreigntableid,
     935             :                                   List **columns)
     936             : {
     937             :     ForeignTable *table;
     938             :     ListCell   *lc;
     939             :     Relation    rel;
     940             :     TupleDesc   tupleDesc;
     941             :     int         attidx;
     942          80 :     Bitmapset  *attrs_used = NULL;
     943          80 :     bool        has_wholerow = false;
     944             :     int         numattrs;
     945             :     int         i;
     946             : 
     947          80 :     *columns = NIL;             /* default result */
     948             : 
     949             :     /*
     950             :      * Check format of the file.  If binary format, this is irrelevant.
     951             :      */
     952          80 :     table = GetForeignTable(foreigntableid);
     953          80 :     foreach(lc, table->options)
     954             :     {
     955          80 :         DefElem    *def = (DefElem *) lfirst(lc);
     956             : 
     957          80 :         if (strcmp(def->defname, "format") == 0)
     958             :         {
     959          80 :             char       *format = defGetString(def);
     960             : 
     961          80 :             if (strcmp(format, "binary") == 0)
     962           0 :                 return false;
     963          80 :             break;
     964             :         }
     965             :     }
     966             : 
     967             :     /* Collect all the attributes needed for joins or final output. */
     968          80 :     pull_varattnos((Node *) baserel->reltarget->exprs, baserel->relid,
     969             :                    &attrs_used);
     970             : 
     971             :     /* Add all the attributes used by restriction clauses. */
     972          96 :     foreach(lc, baserel->baserestrictinfo)
     973             :     {
     974          16 :         RestrictInfo *rinfo = (RestrictInfo *) lfirst(lc);
     975             : 
     976          16 :         pull_varattnos((Node *) rinfo->clause, baserel->relid,
     977             :                        &attrs_used);
     978             :     }
     979             : 
     980             :     /* Convert attribute numbers to column names. */
     981          80 :     rel = table_open(foreigntableid, AccessShareLock);
     982          80 :     tupleDesc = RelationGetDescr(rel);
     983             : 
     984          80 :     attidx = -1;
     985         254 :     while ((attidx = bms_next_member(attrs_used, attidx)) >= 0)
     986             :     {
     987             :         /* attidx is zero-based, attnum is the normal attribute number */
     988         182 :         AttrNumber  attnum = attidx + FirstLowInvalidHeapAttributeNumber;
     989             : 
     990         182 :         if (attnum == 0)
     991             :         {
     992           8 :             has_wholerow = true;
     993           8 :             break;
     994             :         }
     995             : 
     996             :         /* Ignore system attributes. */
     997         174 :         if (attnum < 0)
     998          26 :             continue;
     999             : 
    1000             :         /* Get user attributes. */
    1001         148 :         if (attnum > 0)
    1002             :         {
    1003         148 :             Form_pg_attribute attr = TupleDescAttr(tupleDesc, attnum - 1);
    1004         148 :             char       *attname = NameStr(attr->attname);
    1005             : 
    1006             :             /* Skip dropped attributes (probably shouldn't see any here). */
    1007         148 :             if (attr->attisdropped)
    1008           0 :                 continue;
    1009             : 
    1010             :             /*
    1011             :              * Skip generated columns (COPY won't accept them in the column
    1012             :              * list)
    1013             :              */
    1014         148 :             if (attr->attgenerated)
    1015           2 :                 continue;
    1016         146 :             *columns = lappend(*columns, makeString(pstrdup(attname)));
    1017             :         }
    1018             :     }
    1019             : 
    1020             :     /* Count non-dropped user attributes while we have the tupdesc. */
    1021          80 :     numattrs = 0;
    1022         252 :     for (i = 0; i < tupleDesc->natts; i++)
    1023             :     {
    1024         172 :         Form_pg_attribute attr = TupleDescAttr(tupleDesc, i);
    1025             : 
    1026         172 :         if (attr->attisdropped)
    1027           0 :             continue;
    1028         172 :         numattrs++;
    1029             :     }
    1030             : 
    1031          80 :     table_close(rel, AccessShareLock);
    1032             : 
    1033             :     /* If there's a whole-row reference, fail: we need all the columns. */
    1034          80 :     if (has_wholerow)
    1035             :     {
    1036           8 :         *columns = NIL;
    1037           8 :         return false;
    1038             :     }
    1039             : 
    1040             :     /* If all the user attributes are needed, fail. */
    1041          72 :     if (numattrs == list_length(*columns))
    1042             :     {
    1043          64 :         *columns = NIL;
    1044          64 :         return false;
    1045             :     }
    1046             : 
    1047           8 :     return true;
    1048             : }
    1049             : 
    1050             : /*
    1051             :  * Estimate size of a foreign table.
    1052             :  *
    1053             :  * The main result is returned in baserel->rows.  We also set
    1054             :  * fdw_private->pages and fdw_private->ntuples for later use in the cost
    1055             :  * calculation.
    1056             :  */
    1057             : static void
    1058          80 : estimate_size(PlannerInfo *root, RelOptInfo *baserel,
    1059             :               FileFdwPlanState *fdw_private)
    1060             : {
    1061             :     struct stat stat_buf;
    1062             :     BlockNumber pages;
    1063             :     double      ntuples;
    1064             :     double      nrows;
    1065             : 
    1066             :     /*
    1067             :      * Get size of the file.  It might not be there at plan time, though, in
    1068             :      * which case we have to use a default estimate.  We also have to fall
    1069             :      * back to the default if using a program as the input.
    1070             :      */
    1071          80 :     if (fdw_private->is_program || stat(fdw_private->filename, &stat_buf) < 0)
    1072           0 :         stat_buf.st_size = 10 * BLCKSZ;
    1073             : 
    1074             :     /*
    1075             :      * Convert size to pages for use in I/O cost estimate later.
    1076             :      */
    1077          80 :     pages = (stat_buf.st_size + (BLCKSZ - 1)) / BLCKSZ;
    1078          80 :     if (pages < 1)
    1079           0 :         pages = 1;
    1080          80 :     fdw_private->pages = pages;
    1081             : 
    1082             :     /*
    1083             :      * Estimate the number of tuples in the file.
    1084             :      */
    1085          80 :     if (baserel->tuples >= 0 && baserel->pages > 0)
    1086           0 :     {
    1087             :         /*
    1088             :          * We have # of pages and # of tuples from pg_class (that is, from a
    1089             :          * previous ANALYZE), so compute a tuples-per-page estimate and scale
    1090             :          * that by the current file size.
    1091             :          */
    1092             :         double      density;
    1093             : 
    1094           0 :         density = baserel->tuples / (double) baserel->pages;
    1095           0 :         ntuples = clamp_row_est(density * (double) pages);
    1096             :     }
    1097             :     else
    1098             :     {
    1099             :         /*
    1100             :          * Otherwise we have to fake it.  We back into this estimate using the
    1101             :          * planner's idea of the relation width; which is bogus if not all
    1102             :          * columns are being read, not to mention that the text representation
    1103             :          * of a row probably isn't the same size as its internal
    1104             :          * representation.  Possibly we could do something better, but the
    1105             :          * real answer to anyone who complains is "ANALYZE" ...
    1106             :          */
    1107             :         int         tuple_width;
    1108             : 
    1109          80 :         tuple_width = MAXALIGN(baserel->reltarget->width) +
    1110             :             MAXALIGN(SizeofHeapTupleHeader);
    1111          80 :         ntuples = clamp_row_est((double) stat_buf.st_size /
    1112          80 :                                 (double) tuple_width);
    1113             :     }
    1114          80 :     fdw_private->ntuples = ntuples;
    1115             : 
    1116             :     /*
    1117             :      * Now estimate the number of rows returned by the scan after applying the
    1118             :      * baserestrictinfo quals.
    1119             :      */
    1120          80 :     nrows = ntuples *
    1121          80 :         clauselist_selectivity(root,
    1122             :                                baserel->baserestrictinfo,
    1123             :                                0,
    1124             :                                JOIN_INNER,
    1125             :                                NULL);
    1126             : 
    1127          80 :     nrows = clamp_row_est(nrows);
    1128             : 
    1129             :     /* Save the output-rows estimate for the planner */
    1130          80 :     baserel->rows = nrows;
    1131          80 : }
    1132             : 
    1133             : /*
    1134             :  * Estimate costs of scanning a foreign table.
    1135             :  *
    1136             :  * Results are returned in *startup_cost and *total_cost.
    1137             :  */
    1138             : static void
    1139          80 : estimate_costs(PlannerInfo *root, RelOptInfo *baserel,
    1140             :                FileFdwPlanState *fdw_private,
    1141             :                Cost *startup_cost, Cost *total_cost)
    1142             : {
    1143          80 :     BlockNumber pages = fdw_private->pages;
    1144          80 :     double      ntuples = fdw_private->ntuples;
    1145          80 :     Cost        run_cost = 0;
    1146             :     Cost        cpu_per_tuple;
    1147             : 
    1148             :     /*
    1149             :      * We estimate costs almost the same way as cost_seqscan(), thus assuming
    1150             :      * that I/O costs are equivalent to a regular table file of the same size.
    1151             :      * However, we take per-tuple CPU costs as 10x of a seqscan, to account
    1152             :      * for the cost of parsing records.
    1153             :      *
    1154             :      * In the case of a program source, this calculation is even more divorced
    1155             :      * from reality, but we have no good alternative; and it's not clear that
    1156             :      * the numbers we produce here matter much anyway, since there's only one
    1157             :      * access path for the rel.
    1158             :      */
    1159          80 :     run_cost += seq_page_cost * pages;
    1160             : 
    1161          80 :     *startup_cost = baserel->baserestrictcost.startup;
    1162          80 :     cpu_per_tuple = cpu_tuple_cost * 10 + baserel->baserestrictcost.per_tuple;
    1163          80 :     run_cost += cpu_per_tuple * ntuples;
    1164          80 :     *total_cost = *startup_cost + run_cost;
    1165          80 : }
    1166             : 
    1167             : /*
    1168             :  * file_acquire_sample_rows -- acquire a random sample of rows from the table
    1169             :  *
    1170             :  * Selected rows are returned in the caller-allocated array rows[],
    1171             :  * which must have at least targrows entries.
    1172             :  * The actual number of rows selected is returned as the function result.
    1173             :  * We also count the total number of rows in the file and return it into
    1174             :  * *totalrows.  Rows skipped due to on_error = 'ignore' are not included
    1175             :  * in this count.  Note that *totaldeadrows is always set to 0.
    1176             :  *
    1177             :  * Note that the returned list of rows is not always in order by physical
    1178             :  * position in the file.  Therefore, correlation estimates derived later
    1179             :  * may be meaningless, but it's OK because we don't use the estimates
    1180             :  * currently (the planner only pays attention to correlation for indexscans).
    1181             :  */
    1182             : static int
    1183           2 : file_acquire_sample_rows(Relation onerel, int elevel,
    1184             :                          HeapTuple *rows, int targrows,
    1185             :                          double *totalrows, double *totaldeadrows)
    1186             : {
    1187           2 :     int         numrows = 0;
    1188           2 :     double      rowstoskip = -1;    /* -1 means not set yet */
    1189             :     ReservoirStateData rstate;
    1190             :     TupleDesc   tupDesc;
    1191             :     Datum      *values;
    1192             :     bool       *nulls;
    1193             :     bool        found;
    1194             :     char       *filename;
    1195             :     bool        is_program;
    1196             :     List       *options;
    1197             :     CopyFromState cstate;
    1198             :     ErrorContextCallback errcallback;
    1199           2 :     MemoryContext oldcontext = CurrentMemoryContext;
    1200             :     MemoryContext tupcontext;
    1201             : 
    1202             :     Assert(onerel);
    1203             :     Assert(targrows > 0);
    1204             : 
    1205           2 :     tupDesc = RelationGetDescr(onerel);
    1206           2 :     values = (Datum *) palloc(tupDesc->natts * sizeof(Datum));
    1207           2 :     nulls = (bool *) palloc(tupDesc->natts * sizeof(bool));
    1208             : 
    1209             :     /* Fetch options of foreign table */
    1210           2 :     fileGetOptions(RelationGetRelid(onerel), &filename, &is_program, &options);
    1211             : 
    1212             :     /*
    1213             :      * Create CopyState from FDW options.
    1214             :      */
    1215           2 :     cstate = BeginCopyFrom(NULL, onerel, NULL, filename, is_program, NULL, NIL,
    1216             :                            options);
    1217             : 
    1218             :     /*
    1219             :      * Use per-tuple memory context to prevent leak of memory used to read
    1220             :      * rows from the file with Copy routines.
    1221             :      */
    1222           2 :     tupcontext = AllocSetContextCreate(CurrentMemoryContext,
    1223             :                                        "file_fdw temporary context",
    1224             :                                        ALLOCSET_DEFAULT_SIZES);
    1225             : 
    1226             :     /* Prepare for sampling rows */
    1227           2 :     reservoir_init_selection_state(&rstate, targrows);
    1228             : 
    1229             :     /* Set up callback to identify error line number. */
    1230           2 :     errcallback.callback = CopyFromErrorCallback;
    1231           2 :     errcallback.arg = cstate;
    1232           2 :     errcallback.previous = error_context_stack;
    1233           2 :     error_context_stack = &errcallback;
    1234             : 
    1235           2 :     *totalrows = 0;
    1236           2 :     *totaldeadrows = 0;
    1237             :     for (;;)
    1238             :     {
    1239             :         /* Check for user-requested abort or sleep */
    1240          10 :         vacuum_delay_point();
    1241             : 
    1242             :         /* Fetch next row */
    1243          10 :         MemoryContextReset(tupcontext);
    1244          10 :         MemoryContextSwitchTo(tupcontext);
    1245             : 
    1246          10 :         found = NextCopyFrom(cstate, NULL, values, nulls);
    1247             : 
    1248          10 :         MemoryContextSwitchTo(oldcontext);
    1249             : 
    1250          10 :         if (!found)
    1251           2 :             break;
    1252             : 
    1253           8 :         if (cstate->opts.on_error == COPY_ON_ERROR_IGNORE &&
    1254           8 :             cstate->escontext->error_occurred)
    1255             :         {
    1256             :             /*
    1257             :              * Soft error occurred, skip this tuple and just make
    1258             :              * ErrorSaveContext ready for the next NextCopyFrom. Since we
    1259             :              * don't set details_wanted and error_data is not to be filled,
    1260             :              * just resetting error_occurred is enough.
    1261             :              */
    1262           4 :             cstate->escontext->error_occurred = false;
    1263             : 
    1264             :             /* Repeat NextCopyFrom() until no soft error occurs */
    1265           4 :             continue;
    1266             :         }
    1267             : 
    1268             :         /*
    1269             :          * The first targrows sample rows are simply copied into the
    1270             :          * reservoir.  Then we start replacing tuples in the sample until we
    1271             :          * reach the end of the relation. This algorithm is from Jeff Vitter's
    1272             :          * paper (see more info in commands/analyze.c).
    1273             :          */
    1274           4 :         if (numrows < targrows)
    1275             :         {
    1276           4 :             rows[numrows++] = heap_form_tuple(tupDesc, values, nulls);
    1277             :         }
    1278             :         else
    1279             :         {
    1280             :             /*
    1281             :              * t in Vitter's paper is the number of records already processed.
    1282             :              * If we need to compute a new S value, we must use the
    1283             :              * not-yet-incremented value of totalrows as t.
    1284             :              */
    1285           0 :             if (rowstoskip < 0)
    1286           0 :                 rowstoskip = reservoir_get_next_S(&rstate, *totalrows, targrows);
    1287             : 
    1288           0 :             if (rowstoskip <= 0)
    1289             :             {
    1290             :                 /*
    1291             :                  * Found a suitable tuple, so save it, replacing one old tuple
    1292             :                  * at random
    1293             :                  */
    1294           0 :                 int         k = (int) (targrows * sampler_random_fract(&rstate.randstate));
    1295             : 
    1296             :                 Assert(k >= 0 && k < targrows);
    1297           0 :                 heap_freetuple(rows[k]);
    1298           0 :                 rows[k] = heap_form_tuple(tupDesc, values, nulls);
    1299             :             }
    1300             : 
    1301           0 :             rowstoskip -= 1;
    1302             :         }
    1303             : 
    1304           4 :         *totalrows += 1;
    1305             :     }
    1306             : 
    1307             :     /* Remove error callback. */
    1308           2 :     error_context_stack = errcallback.previous;
    1309             : 
    1310             :     /* Clean up. */
    1311           2 :     MemoryContextDelete(tupcontext);
    1312             : 
    1313           2 :     if (cstate->opts.on_error == COPY_ON_ERROR_IGNORE &&
    1314           2 :         cstate->num_errors > 0 &&
    1315           2 :         cstate->opts.log_verbosity >= COPY_LOG_VERBOSITY_DEFAULT)
    1316           0 :         ereport(NOTICE,
    1317             :                 errmsg_plural("%llu row was skipped due to data type incompatibility",
    1318             :                               "%llu rows were skipped due to data type incompatibility",
    1319             :                               (unsigned long long) cstate->num_errors,
    1320             :                               (unsigned long long) cstate->num_errors));
    1321             : 
    1322           2 :     EndCopyFrom(cstate);
    1323             : 
    1324           2 :     pfree(values);
    1325           2 :     pfree(nulls);
    1326             : 
    1327             :     /*
    1328             :      * Emit some interesting relation info
    1329             :      */
    1330           2 :     ereport(elevel,
    1331             :             (errmsg("\"%s\": file contains %.0f rows; "
    1332             :                     "%d rows in sample",
    1333             :                     RelationGetRelationName(onerel),
    1334             :                     *totalrows, numrows)));
    1335             : 
    1336           2 :     return numrows;
    1337             : }

Generated by: LCOV version 1.14