codekingpro/portable-devtools
115k
1<?xml version="1.0" encoding="UTF-8" standalone="no"?>2<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Transitional//EN" "http://www.w3.org/TR/xhtml1/DTD/xhtml1-transitional.dtd"><html xmlns="http://www.w3.org/1999/xhtml"><head><meta http-equiv="Content-Type" content="text/html; charset=UTF-8" /><title>59.2. Foreign Data Wrapper Callback Routines</title><link rel="stylesheet" type="text/css" href="stylesheet.css" /><link rev="made" href="pgsql-docs@lists.postgresql.org" /><meta name="generator" content="DocBook XSL Stylesheets Vsnapshot" /><link rel="prev" href="fdw-functions.html" title="59.1. Foreign Data Wrapper Functions" /><link rel="next" href="fdw-helpers.html" title="59.3. Foreign Data Wrapper Helper Functions" /></head><body id="docContent" class="container-fluid col-10"><div class="navheader"><table width="100%" summary="Navigation header"><tr><th colspan="5" align="center">59.2. Foreign Data Wrapper Callback Routines</th></tr><tr><td width="10%" align="left"><a accesskey="p" href="fdw-functions.html" title="59.1. Foreign Data Wrapper Functions">Prev</a> </td><td width="10%" align="left"><a accesskey="u" href="fdwhandler.html" title="Chapter 59. Writing a Foreign Data Wrapper">Up</a></td><th width="60%" align="center">Chapter 59. Writing a Foreign Data Wrapper</th><td width="10%" align="right"><a accesskey="h" href="index.html" title="PostgreSQL 16.3 Documentation">Home</a></td><td width="10%" align="right"> <a accesskey="n" href="fdw-helpers.html" title="59.3. Foreign Data Wrapper Helper Functions">Next</a></td></tr></table><hr /></div><div class="sect1" id="FDW-CALLBACKS"><div class="titlepage"><div><div><h2 class="title" style="clear: both">59.2. Foreign Data Wrapper Callback Routines <a href="#FDW-CALLBACKS" class="id_link">#</a></h2></div></div></div><div class="toc"><dl class="toc"><dt><span class="sect2"><a href="fdw-callbacks.html#FDW-CALLBACKS-SCAN">59.2.1. FDW Routines for Scanning Foreign Tables</a></span></dt><dt><span class="sect2"><a href="fdw-callbacks.html#FDW-CALLBACKS-JOIN-SCAN">59.2.2. FDW Routines for Scanning Foreign Joins</a></span></dt><dt><span class="sect2"><a href="fdw-callbacks.html#FDW-CALLBACKS-UPPER-PLANNING">59.2.3. FDW Routines for Planning Post-Scan/Join Processing</a></span></dt><dt><span class="sect2"><a href="fdw-callbacks.html#FDW-CALLBACKS-UPDATE">59.2.4. FDW Routines for Updating Foreign Tables</a></span></dt><dt><span class="sect2"><a href="fdw-callbacks.html#FDW-CALLBACKS-TRUNCATE">59.2.5. FDW Routines for <code class="command">TRUNCATE</code></a></span></dt><dt><span class="sect2"><a href="fdw-callbacks.html#FDW-CALLBACKS-ROW-LOCKING">59.2.6. FDW Routines for Row Locking</a></span></dt><dt><span class="sect2"><a href="fdw-callbacks.html#FDW-CALLBACKS-EXPLAIN">59.2.7. FDW Routines for <code class="command">EXPLAIN</code></a></span></dt><dt><span class="sect2"><a href="fdw-callbacks.html#FDW-CALLBACKS-ANALYZE">59.2.8. FDW Routines for <code class="command">ANALYZE</code></a></span></dt><dt><span class="sect2"><a href="fdw-callbacks.html#FDW-CALLBACKS-IMPORT">59.2.9. FDW Routines for <code class="command">IMPORT FOREIGN SCHEMA</code></a></span></dt><dt><span class="sect2"><a href="fdw-callbacks.html#FDW-CALLBACKS-PARALLEL">59.2.10. FDW Routines for Parallel Execution</a></span></dt><dt><span class="sect2"><a href="fdw-callbacks.html#FDW-CALLBACKS-ASYNC">59.2.11. FDW Routines for Asynchronous Execution</a></span></dt><dt><span class="sect2"><a href="fdw-callbacks.html#FDW-CALLBACKS-REPARAMETERIZE-PATHS">59.2.12. FDW Routines for Reparameterization of Paths</a></span></dt></dl></div><p>3 The FDW handler function returns a palloc'd <code class="structname">FdwRoutine</code>4 struct containing pointers to the callback functions described below.5 The scan-related functions are required, the rest are optional.6 </p><p>7 The <code class="structname">FdwRoutine</code> struct type is declared in8 <code class="filename">src/include/foreign/fdwapi.h</code>, which see for additional9 details.10 </p><div class="sect2" id="FDW-CALLBACKS-SCAN"><div class="titlepage"><div><div><h3 class="title">59.2.1. FDW Routines for Scanning Foreign Tables <a href="#FDW-CALLBACKS-SCAN" class="id_link">#</a></h3></div></div></div><p>11</p><pre class="programlisting">12void13GetForeignRelSize(PlannerInfo *root,14 RelOptInfo *baserel,15 Oid foreigntableid);16</pre><p>17 18 Obtain relation size estimates for a foreign table. This is called19 at the beginning of planning for a query that scans a foreign table.20 <code class="literal">root</code> is the planner's global information about the query;21 <code class="literal">baserel</code> is the planner's information about this table; and22 <code class="literal">foreigntableid</code> is the <code class="structname">pg_class</code> OID of the23 foreign table. (<code class="literal">foreigntableid</code> could be obtained from the24 planner data structures, but it's passed explicitly to save effort.)25 </p><p>26 This function should update <code class="literal">baserel->rows</code> to be the27 expected number of rows returned by the table scan, after accounting for28 the filtering done by the restriction quals. The initial value of29 <code class="literal">baserel->rows</code> is just a constant default estimate, which30 should be replaced if at all possible. The function may also choose to31 update <code class="literal">baserel->width</code> if it can compute a better estimate32 of the average result row width.33 (The initial value is based on column data types and on column34 average-width values measured by the last <code class="command">ANALYZE</code>.)35 Also, this function may update <code class="literal">baserel->tuples</code> if36 it can compute a better estimate of the foreign table's total row count.37 (The initial value is38 from <code class="structname">pg_class</code>.<code class="structfield">reltuples</code>39 which represents the total row count seen by the40 last <code class="command">ANALYZE</code>; it will be <code class="literal">-1</code> if41 no <code class="command">ANALYZE</code> has been done on this foreign table.)42 </p><p>43 See <a class="xref" href="fdw-planning.html" title="59.4. Foreign Data Wrapper Query Planning">Section 59.4</a> for additional information.44 </p><p>45</p><pre class="programlisting">46void47GetForeignPaths(PlannerInfo *root,48 RelOptInfo *baserel,49 Oid foreigntableid);50</pre><p>51 52 Create possible access paths for a scan on a foreign table.53 This is called during query planning.54 The parameters are the same as for <code class="function">GetForeignRelSize</code>,55 which has already been called.56 </p><p>57 This function must generate at least one access path58 (<code class="structname">ForeignPath</code> node) for a scan on the foreign table and59 must call <code class="function">add_path</code> to add each such path to60 <code class="literal">baserel->pathlist</code>. It's recommended to use61 <code class="function">create_foreignscan_path</code> to build the62 <code class="structname">ForeignPath</code> nodes. The function can generate multiple63 access paths, e.g., a path which has valid <code class="literal">pathkeys</code> to64 represent a pre-sorted result. Each access path must contain cost65 estimates, and can contain any FDW-private information that is needed to66 identify the specific scan method intended.67 </p><p>68 See <a class="xref" href="fdw-planning.html" title="59.4. Foreign Data Wrapper Query Planning">Section 59.4</a> for additional information.69 </p><p>70</p><pre class="programlisting">71ForeignScan *72GetForeignPlan(PlannerInfo *root,73 RelOptInfo *baserel,74 Oid foreigntableid,75 ForeignPath *best_path,76 List *tlist,77 List *scan_clauses,78 Plan *outer_plan);79</pre><p>80 81 Create a <code class="structname">ForeignScan</code> plan node from the selected foreign82 access path. This is called at the end of query planning.83 The parameters are as for <code class="function">GetForeignRelSize</code>, plus84 the selected <code class="structname">ForeignPath</code> (previously produced by85 <code class="function">GetForeignPaths</code>, <code class="function">GetForeignJoinPaths</code>,86 or <code class="function">GetForeignUpperPaths</code>),87 the target list to be emitted by the plan node,88 the restriction clauses to be enforced by the plan node,89 and the outer subplan of the <code class="structname">ForeignScan</code>,90 which is used for rechecks performed by <code class="function">RecheckForeignScan</code>.91 (If the path is for a join rather than a base92 relation, <code class="literal">foreigntableid</code> is <code class="literal">InvalidOid</code>.)93 </p><p>94 This function must create and return a <code class="structname">ForeignScan</code> plan95 node; it's recommended to use <code class="function">make_foreignscan</code> to build the96 <code class="structname">ForeignScan</code> node.97 </p><p>98 See <a class="xref" href="fdw-planning.html" title="59.4. Foreign Data Wrapper Query Planning">Section 59.4</a> for additional information.99 </p><p>100</p><pre class="programlisting">101void102BeginForeignScan(ForeignScanState *node,103 int eflags);104</pre><p>105 106 Begin executing a foreign scan. This is called during executor startup.107 It should perform any initialization needed before the scan can start,108 but not start executing the actual scan (that should be done upon the109 first call to <code class="function">IterateForeignScan</code>).110 The <code class="structname">ForeignScanState</code> node has already been created, but111 its <code class="structfield">fdw_state</code> field is still NULL. Information about112 the table to scan is accessible through the113 <code class="structname">ForeignScanState</code> node (in particular, from the underlying114 <code class="structname">ForeignScan</code> plan node, which contains any FDW-private115 information provided by <code class="function">GetForeignPlan</code>).116 <code class="literal">eflags</code> contains flag bits describing the executor's117 operating mode for this plan node.118 </p><p>119 Note that when <code class="literal">(eflags & EXEC_FLAG_EXPLAIN_ONLY)</code> is120 true, this function should not perform any externally-visible actions;121 it should only do the minimum required to make the node state valid122 for <code class="function">ExplainForeignScan</code> and <code class="function">EndForeignScan</code>.123 </p><p>124</p><pre class="programlisting">125TupleTableSlot *126IterateForeignScan(ForeignScanState *node);127</pre><p>128 129 Fetch one row from the foreign source, returning it in a tuple table slot130 (the node's <code class="structfield">ScanTupleSlot</code> should be used for this131 purpose). Return NULL if no more rows are available. The tuple table132 slot infrastructure allows either a physical or virtual tuple to be133 returned; in most cases the latter choice is preferable from a134 performance standpoint. Note that this is called in a short-lived memory135 context that will be reset between invocations. Create a memory context136 in <code class="function">BeginForeignScan</code> if you need longer-lived storage, or use137 the <code class="structfield">es_query_cxt</code> of the node's <code class="structname">EState</code>.138 </p><p>139 The rows returned must match the <code class="structfield">fdw_scan_tlist</code> target140 list if one was supplied, otherwise they must match the row type of the141 foreign table being scanned. If you choose to optimize away fetching142 columns that are not needed, you should insert nulls in those column143 positions, or else generate a <code class="structfield">fdw_scan_tlist</code> list with144 those columns omitted.145 </p><p>146 Note that <span class="productname">PostgreSQL</span>'s executor doesn't care147 whether the rows returned violate any constraints that were defined on148 the foreign table — but the planner does care, and may optimize149 queries incorrectly if there are rows visible in the foreign table that150 do not satisfy a declared constraint. If a constraint is violated when151 the user has declared that the constraint should hold true, it may be152 appropriate to raise an error (just as you would need to do in the case153 of a data type mismatch).154 </p><p>155</p><pre class="programlisting">156void157ReScanForeignScan(ForeignScanState *node);158</pre><p>159 160 Restart the scan from the beginning. Note that any parameters the161 scan depends on may have changed value, so the new scan does not162 necessarily return exactly the same rows.163 </p><p>164</p><pre class="programlisting">165void166EndForeignScan(ForeignScanState *node);167</pre><p>168 169 End the scan and release resources. It is normally not important170 to release palloc'd memory, but for example open files and connections171 to remote servers should be cleaned up.172 </p></div><div class="sect2" id="FDW-CALLBACKS-JOIN-SCAN"><div class="titlepage"><div><div><h3 class="title">59.2.2. FDW Routines for Scanning Foreign Joins <a href="#FDW-CALLBACKS-JOIN-SCAN" class="id_link">#</a></h3></div></div></div><p>173 If an FDW supports performing foreign joins remotely (rather than174 by fetching both tables' data and doing the join locally), it should175 provide this callback function:176 </p><p>177</p><pre class="programlisting">178void179GetForeignJoinPaths(PlannerInfo *root,180 RelOptInfo *joinrel,181 RelOptInfo *outerrel,182 RelOptInfo *innerrel,183 JoinType jointype,184 JoinPathExtraData *extra);185</pre><p>186 Create possible access paths for a join of two (or more) foreign tables187 that all belong to the same foreign server. This optional188 function is called during query planning. As189 with <code class="function">GetForeignPaths</code>, this function should190 generate <code class="structname">ForeignPath</code> path(s) for the191 supplied <code class="literal">joinrel</code>192 (use <code class="function">create_foreign_join_path</code> to build them),193 and call <code class="function">add_path</code> to add these194 paths to the set of paths considered for the join. But unlike195 <code class="function">GetForeignPaths</code>, it is not necessary that this function196 succeed in creating at least one path, since paths involving local197 joining are always possible.198 </p><p>199 Note that this function will be invoked repeatedly for the same join200 relation, with different combinations of inner and outer relations; it is201 the responsibility of the FDW to minimize duplicated work.202 </p><p>203 If a <code class="structname">ForeignPath</code> path is chosen for the join, it will204 represent the entire join process; paths generated for the component205 tables and subsidiary joins will not be used. Subsequent processing of206 the join path proceeds much as it does for a path scanning a single207 foreign table. One difference is that the <code class="structfield">scanrelid</code> of208 the resulting <code class="structname">ForeignScan</code> plan node should be set to zero,209 since there is no single relation that it represents; instead,210 the <code class="structfield">fs_relids</code> field of the <code class="structname">ForeignScan</code>211 node represents the set of relations that were joined. (The latter field212 is set up automatically by the core planner code, and need not be filled213 by the FDW.) Another difference is that, because the column list for a214 remote join cannot be found from the system catalogs, the FDW must215 fill <code class="structfield">fdw_scan_tlist</code> with an appropriate list216 of <code class="structfield">TargetEntry</code> nodes, representing the set of columns217 it will supply at run time in the tuples it returns.218 </p><div class="note"><h3 class="title">Note</h3><p>219 Beginning with <span class="productname">PostgreSQL</span> 16,220 <code class="structfield">fs_relids</code> includes the rangetable indexes221 of outer joins, if any were involved in this join. The new field222 <code class="structfield">fs_base_relids</code> includes only base223 relation indexes, and thus224 mimics <code class="structfield">fs_relids</code>'s old semantics.225 </p></div><p>226 See <a class="xref" href="fdw-planning.html" title="59.4. Foreign Data Wrapper Query Planning">Section 59.4</a> for additional information.227 </p></div><div class="sect2" id="FDW-CALLBACKS-UPPER-PLANNING"><div class="titlepage"><div><div><h3 class="title">59.2.3. FDW Routines for Planning Post-Scan/Join Processing <a href="#FDW-CALLBACKS-UPPER-PLANNING" class="id_link">#</a></h3></div></div></div><p>228 If an FDW supports performing remote post-scan/join processing, such as229 remote aggregation, it should provide this callback function:230 </p><p>231</p><pre class="programlisting">232void233GetForeignUpperPaths(PlannerInfo *root,234 UpperRelationKind stage,235 RelOptInfo *input_rel,236 RelOptInfo *output_rel,237 void *extra);238</pre><p>239 Create possible access paths for <em class="firstterm">upper relation</em> processing,240 which is the planner's term for all post-scan/join query processing, such241 as aggregation, window functions, sorting, and table updates. This242 optional function is called during query planning. Currently, it is243 called only if all base relation(s) involved in the query belong to the244 same FDW. This function should generate <code class="structname">ForeignPath</code>245 path(s) for any post-scan/join processing that the FDW knows how to246 perform remotely247 (use <code class="function">create_foreign_upper_path</code> to build them),248 and call <code class="function">add_path</code> to add these paths to249 the indicated upper relation. As with <code class="function">GetForeignJoinPaths</code>,250 it is not necessary that this function succeed in creating any paths,251 since paths involving local processing are always possible.252 </p><p>253 The <code class="literal">stage</code> parameter identifies which post-scan/join step is254 currently being considered. <code class="literal">output_rel</code> is the upper relation255 that should receive paths representing computation of this step,256 and <code class="literal">input_rel</code> is the relation representing the input to this257 step. The <code class="literal">extra</code> parameter provides additional details,258 currently, it is set only for <code class="literal">UPPERREL_PARTIAL_GROUP_AGG</code>259 or <code class="literal">UPPERREL_GROUP_AGG</code>, in which case it points to a260 <code class="literal">GroupPathExtraData</code> structure;261 or for <code class="literal">UPPERREL_FINAL</code>, in which case it points to a262 <code class="literal">FinalPathExtraData</code> structure.263 (Note that <code class="structname">ForeignPath</code> paths added264 to <code class="literal">output_rel</code> would typically not have any direct dependency265 on paths of the <code class="literal">input_rel</code>, since their processing is expected266 to be done externally. However, examining paths previously generated for267 the previous processing step can be useful to avoid redundant planning268 work.)269 </p><p>270 See <a class="xref" href="fdw-planning.html" title="59.4. Foreign Data Wrapper Query Planning">Section 59.4</a> for additional information.271 </p></div><div class="sect2" id="FDW-CALLBACKS-UPDATE"><div class="titlepage"><div><div><h3 class="title">59.2.4. FDW Routines for Updating Foreign Tables <a href="#FDW-CALLBACKS-UPDATE" class="id_link">#</a></h3></div></div></div><p>272 If an FDW supports writable foreign tables, it should provide273 some or all of the following callback functions depending on274 the needs and capabilities of the FDW:275 </p><p>276</p><pre class="programlisting">277void278AddForeignUpdateTargets(PlannerInfo *root,279 Index rtindex,280 RangeTblEntry *target_rte,281 Relation target_relation);282</pre><p>283 284 <code class="command">UPDATE</code> and <code class="command">DELETE</code> operations are performed285 against rows previously fetched by the table-scanning functions. The286 FDW may need extra information, such as a row ID or the values of287 primary-key columns, to ensure that it can identify the exact row to288 update or delete. To support that, this function can add extra hidden,289 or <span class="quote">“<span class="quote">junk</span>”</span>, target columns to the list of columns that are to be290 retrieved from the foreign table during an <code class="command">UPDATE</code> or291 <code class="command">DELETE</code>.292 </p><p>293 To do that, construct a <code class="structname">Var</code> representing294 an extra value you need, and pass it295 to <code class="function">add_row_identity_var</code>, along with a name for296 the junk column. (You can do this more than once if several columns297 are needed.) You must choose a distinct junk column name for each298 different <code class="structname">Var</code> you need, except299 that <code class="structname">Var</code>s that are identical except for300 the <code class="structfield">varno</code> field can and should share a301 column name.302 The core system uses the junk column names303 <code class="literal">tableoid</code> for a304 table's <code class="structfield">tableoid</code> column,305 <code class="literal">ctid</code>306 or <code class="literal">ctid<em class="replaceable"><code>N</code></em></code>307 for <code class="structfield">ctid</code>,308 <code class="literal">wholerow</code>309 for a whole-row <code class="structname">Var</code> marked with310 <code class="structfield">vartype</code> = <code class="type">RECORD</code>,311 and <code class="literal">wholerow<em class="replaceable"><code>N</code></em></code>312 for a whole-row <code class="structname">Var</code> with313 <code class="structfield">vartype</code> equal to the table's declared row type.314 Re-use these names when you can (the planner will combine duplicate315 requests for identical junk columns). If you need another kind of316 junk column besides these, it might be wise to choose a name prefixed317 with your extension name, to avoid conflicts against other FDWs.318 </p><p>319 If the <code class="function">AddForeignUpdateTargets</code> pointer is set to320 <code class="literal">NULL</code>, no extra target expressions are added.321 (This will make it impossible to implement <code class="command">DELETE</code>322 operations, though <code class="command">UPDATE</code> may still be feasible if the FDW323 relies on an unchanging primary key to identify rows.)324 </p><p>325</p><pre class="programlisting">326List *327PlanForeignModify(PlannerInfo *root,328 ModifyTable *plan,329 Index resultRelation,330 int subplan_index);331</pre><p>332 333 Perform any additional planning actions needed for an insert, update, or334 delete on a foreign table. This function generates the FDW-private335 information that will be attached to the <code class="structname">ModifyTable</code> plan336 node that performs the update action. This private information must337 have the form of a <code class="literal">List</code>, and will be delivered to338 <code class="function">BeginForeignModify</code> during the execution stage.339 </p><p>340 <code class="literal">root</code> is the planner's global information about the query.341 <code class="literal">plan</code> is the <code class="structname">ModifyTable</code> plan node, which is342 complete except for the <code class="structfield">fdwPrivLists</code> field.343 <code class="literal">resultRelation</code> identifies the target foreign table by its344 range table index. <code class="literal">subplan_index</code> identifies which target of345 the <code class="structname">ModifyTable</code> plan node this is, counting from zero;346 use this if you want to index into per-target-relation substructures of the347 <code class="literal">plan</code> node.348 </p><p>349 See <a class="xref" href="fdw-planning.html" title="59.4. Foreign Data Wrapper Query Planning">Section 59.4</a> for additional information.350 </p><p>351 If the <code class="function">PlanForeignModify</code> pointer is set to352 <code class="literal">NULL</code>, no additional plan-time actions are taken, and the353 <code class="literal">fdw_private</code> list delivered to354 <code class="function">BeginForeignModify</code> will be NIL.355 </p><p>356</p><pre class="programlisting">357void358BeginForeignModify(ModifyTableState *mtstate,359 ResultRelInfo *rinfo,360 List *fdw_private,361 int subplan_index,362 int eflags);363</pre><p>364 365 Begin executing a foreign table modification operation. This routine is366 called during executor startup. It should perform any initialization367 needed prior to the actual table modifications. Subsequently,368 <code class="function">ExecForeignInsert/ExecForeignBatchInsert</code>,369 <code class="function">ExecForeignUpdate</code> or370 <code class="function">ExecForeignDelete</code> will be called for tuple(s) to be371 inserted, updated, or deleted.372 </p><p>373 <code class="literal">mtstate</code> is the overall state of the374 <code class="structname">ModifyTable</code> plan node being executed; global data about375 the plan and execution state is available via this structure.376 <code class="literal">rinfo</code> is the <code class="structname">ResultRelInfo</code> struct describing377 the target foreign table. (The <code class="structfield">ri_FdwState</code> field of378 <code class="structname">ResultRelInfo</code> is available for the FDW to store any379 private state it needs for this operation.)380 <code class="literal">fdw_private</code> contains the private data generated by381 <code class="function">PlanForeignModify</code>, if any.382 <code class="literal">subplan_index</code> identifies which target of383 the <code class="structname">ModifyTable</code> plan node this is.384 <code class="literal">eflags</code> contains flag bits describing the executor's385 operating mode for this plan node.386 </p><p>387 Note that when <code class="literal">(eflags & EXEC_FLAG_EXPLAIN_ONLY)</code> is388 true, this function should not perform any externally-visible actions;389 it should only do the minimum required to make the node state valid390 for <code class="function">ExplainForeignModify</code> and <code class="function">EndForeignModify</code>.391 </p><p>392 If the <code class="function">BeginForeignModify</code> pointer is set to393 <code class="literal">NULL</code>, no action is taken during executor startup.394 </p><p>395</p><pre class="programlisting">396TupleTableSlot *397ExecForeignInsert(EState *estate,398 ResultRelInfo *rinfo,399 TupleTableSlot *slot,400 TupleTableSlot *planSlot);401</pre><p>402 403 Insert one tuple into the foreign table.404 <code class="literal">estate</code> is global execution state for the query.405 <code class="literal">rinfo</code> is the <code class="structname">ResultRelInfo</code> struct describing406 the target foreign table.407 <code class="literal">slot</code> contains the tuple to be inserted; it will match the408 row-type definition of the foreign table.409 <code class="literal">planSlot</code> contains the tuple that was generated by the410 <code class="structname">ModifyTable</code> plan node's subplan; it differs from411 <code class="literal">slot</code> in possibly containing additional <span class="quote">“<span class="quote">junk</span>”</span>412 columns. (The <code class="literal">planSlot</code> is typically of little interest413 for <code class="command">INSERT</code> cases, but is provided for completeness.)414 </p><p>415 The return value is either a slot containing the data that was actually416 inserted (this might differ from the data supplied, for example as a417 result of trigger actions), or NULL if no row was actually inserted418 (again, typically as a result of triggers). The passed-in419 <code class="literal">slot</code> can be re-used for this purpose.420 </p><p>421 The data in the returned slot is used only if the <code class="command">INSERT</code>422 statement has a <code class="literal">RETURNING</code> clause or involves a view423 <code class="literal">WITH CHECK OPTION</code>; or if the foreign table has424 an <code class="literal">AFTER ROW</code> trigger. Triggers require all columns,425 but the FDW could choose to optimize away returning some or all columns426 depending on the contents of the <code class="literal">RETURNING</code> clause or427 <code class="literal">WITH CHECK OPTION</code> constraints. Regardless, some slot428 must be returned to indicate success, or the query's reported row count429 will be wrong.430 </p><p>431 If the <code class="function">ExecForeignInsert</code> pointer is set to432 <code class="literal">NULL</code>, attempts to insert into the foreign table will fail433 with an error message.434 </p><p>435 Note that this function is also called when inserting routed tuples into436 a foreign-table partition or executing <code class="command">COPY FROM</code> on437 a foreign table, in which case it is called in a different way than it438 is in the <code class="command">INSERT</code> case. See the callback functions439 described below that allow the FDW to support that.440 </p><p>441</p><pre class="programlisting">442TupleTableSlot **443ExecForeignBatchInsert(EState *estate,444 ResultRelInfo *rinfo,445 TupleTableSlot **slots,446 TupleTableSlot **planSlots,447 int *numSlots);448</pre><p>449 450 Insert multiple tuples in bulk into the foreign table.451 The parameters are the same for <code class="function">ExecForeignInsert</code>452 except <code class="literal">slots</code> and <code class="literal">planSlots</code> contain453 multiple tuples and <code class="literal">*numSlots</code> specifies the number of454 tuples in those arrays.455 </p><p>456 The return value is an array of slots containing the data that was457 actually inserted (this might differ from the data supplied, for458 example as a result of trigger actions.)459 The passed-in <code class="literal">slots</code> can be re-used for this purpose.460 The number of successfully inserted tuples is returned in461 <code class="literal">*numSlots</code>.462 </p><p>463 The data in the returned slot is used only if the <code class="command">INSERT</code>464 statement involves a view465 <code class="literal">WITH CHECK OPTION</code>; or if the foreign table has466 an <code class="literal">AFTER ROW</code> trigger. Triggers require all columns,467 but the FDW could choose to optimize away returning some or all columns468 depending on the contents of the469 <code class="literal">WITH CHECK OPTION</code> constraints.470 </p><p>471 If the <code class="function">ExecForeignBatchInsert</code> or472 <code class="function">GetForeignModifyBatchSize</code> pointer is set to473 <code class="literal">NULL</code>, attempts to insert into the foreign table will474 use <code class="function">ExecForeignInsert</code>.475 This function is not used if the <code class="command">INSERT</code> has the476 <code class="literal">RETURNING</code> clause.477 </p><p>478 Note that this function is also called when inserting routed tuples into479 a foreign-table partition or executing <code class="command">COPY FROM</code> on480 a foreign table, in which case it is called in a different way than it481 is in the <code class="command">INSERT</code> case. See the callback functions482 described below that allow the FDW to support that.483 </p><p>484</p><pre class="programlisting">485int486GetForeignModifyBatchSize(ResultRelInfo *rinfo);487</pre><p>488 489 Report the maximum number of tuples that a single490 <code class="function">ExecForeignBatchInsert</code> call can handle for491 the specified foreign table. The executor passes at most492 the given number of tuples to <code class="function">ExecForeignBatchInsert</code>.493 <code class="literal">rinfo</code> is the <code class="structname">ResultRelInfo</code> struct describing494 the target foreign table.495 The FDW is expected to provide a foreign server and/or foreign496 table option for the user to set this value, or some hard-coded value.497 </p><p>498 If the <code class="function">ExecForeignBatchInsert</code> or499 <code class="function">GetForeignModifyBatchSize</code> pointer is set to500 <code class="literal">NULL</code>, attempts to insert into the foreign table will501 use <code class="function">ExecForeignInsert</code>.502 </p><p>503</p><pre class="programlisting">504TupleTableSlot *505ExecForeignUpdate(EState *estate,506 ResultRelInfo *rinfo,507 TupleTableSlot *slot,508 TupleTableSlot *planSlot);509</pre><p>510 511 Update one tuple in the foreign table.512 <code class="literal">estate</code> is global execution state for the query.513 <code class="literal">rinfo</code> is the <code class="structname">ResultRelInfo</code> struct describing514 the target foreign table.515 <code class="literal">slot</code> contains the new data for the tuple; it will match the516 row-type definition of the foreign table.517 <code class="literal">planSlot</code> contains the tuple that was generated by the518 <code class="structname">ModifyTable</code> plan node's subplan. Unlike519 <code class="literal">slot</code>, this tuple contains only the new values for520 columns changed by the query, so do not rely on attribute numbers of the521 foreign table to index into <code class="literal">planSlot</code>.522 Also, <code class="literal">planSlot</code> typically contains523 additional <span class="quote">“<span class="quote">junk</span>”</span> columns. In particular, any junk columns524 that were requested by <code class="function">AddForeignUpdateTargets</code> will525 be available from this slot.526 </p><p>527 The return value is either a slot containing the row as it was actually528 updated (this might differ from the data supplied, for example as a529 result of trigger actions), or NULL if no row was actually updated530 (again, typically as a result of triggers). The passed-in531 <code class="literal">slot</code> can be re-used for this purpose.532 </p><p>533 The data in the returned slot is used only if the <code class="command">UPDATE</code>534 statement has a <code class="literal">RETURNING</code> clause or involves a view535 <code class="literal">WITH CHECK OPTION</code>; or if the foreign table has536 an <code class="literal">AFTER ROW</code> trigger. Triggers require all columns,537 but the FDW could choose to optimize away returning some or all columns538 depending on the contents of the <code class="literal">RETURNING</code> clause or539 <code class="literal">WITH CHECK OPTION</code> constraints. Regardless, some slot540 must be returned to indicate success, or the query's reported row count541 will be wrong.542 </p><p>543 If the <code class="function">ExecForeignUpdate</code> pointer is set to544 <code class="literal">NULL</code>, attempts to update the foreign table will fail545 with an error message.546 </p><p>547</p><pre class="programlisting">548TupleTableSlot *549ExecForeignDelete(EState *estate,550 ResultRelInfo *rinfo,551 TupleTableSlot *slot,552 TupleTableSlot *planSlot);553</pre><p>554 555 Delete one tuple from the foreign table.556 <code class="literal">estate</code> is global execution state for the query.557 <code class="literal">rinfo</code> is the <code class="structname">ResultRelInfo</code> struct describing558 the target foreign table.559 <code class="literal">slot</code> contains nothing useful upon call, but can be used to560 hold the returned tuple.561 <code class="literal">planSlot</code> contains the tuple that was generated by the562 <code class="structname">ModifyTable</code> plan node's subplan; in particular, it will563 carry any junk columns that were requested by564 <code class="function">AddForeignUpdateTargets</code>. The junk column(s) must be used565 to identify the tuple to be deleted.566 </p><p>567 The return value is either a slot containing the row that was deleted,568 or NULL if no row was deleted (typically as a result of triggers). The569 passed-in <code class="literal">slot</code> can be used to hold the tuple to be returned.570 </p><p>571 The data in the returned slot is used only if the <code class="command">DELETE</code>572 query has a <code class="literal">RETURNING</code> clause or the foreign table has573 an <code class="literal">AFTER ROW</code> trigger. Triggers require all columns, but the574 FDW could choose to optimize away returning some or all columns depending575 on the contents of the <code class="literal">RETURNING</code> clause. Regardless, some576 slot must be returned to indicate success, or the query's reported row577 count will be wrong.578 </p><p>579 If the <code class="function">ExecForeignDelete</code> pointer is set to580 <code class="literal">NULL</code>, attempts to delete from the foreign table will fail581 with an error message.582 </p><p>583</p><pre class="programlisting">584void585EndForeignModify(EState *estate,586 ResultRelInfo *rinfo);587</pre><p>588 589 End the table update and release resources. It is normally not important590 to release palloc'd memory, but for example open files and connections591 to remote servers should be cleaned up.592 </p><p>593 If the <code class="function">EndForeignModify</code> pointer is set to594 <code class="literal">NULL</code>, no action is taken during executor shutdown.595 </p><p>596 Tuples inserted into a partitioned table by <code class="command">INSERT</code> or597 <code class="command">COPY FROM</code> are routed to partitions. If an FDW598 supports routable foreign-table partitions, it should also provide the599 following callback functions. These functions are also called when600 <code class="command">COPY FROM</code> is executed on a foreign table.601 </p><p>602</p><pre class="programlisting">603void604BeginForeignInsert(ModifyTableState *mtstate,605 ResultRelInfo *rinfo);606</pre><p>607 608 Begin executing an insert operation on a foreign table. This routine is609 called right before the first tuple is inserted into the foreign table610 in both cases when it is the partition chosen for tuple routing and the611 target specified in a <code class="command">COPY FROM</code> command. It should612 perform any initialization needed prior to the actual insertion.613 Subsequently, <code class="function">ExecForeignInsert</code> or614 <code class="function">ExecForeignBatchInsert</code> will be called for615 tuple(s) to be inserted into the foreign table.616 </p><p>617 <code class="literal">mtstate</code> is the overall state of the618 <code class="structname">ModifyTable</code> plan node being executed; global data about619 the plan and execution state is available via this structure.620 <code class="literal">rinfo</code> is the <code class="structname">ResultRelInfo</code> struct describing621 the target foreign table. (The <code class="structfield">ri_FdwState</code> field of622 <code class="structname">ResultRelInfo</code> is available for the FDW to store any623 private state it needs for this operation.)624 </p><p>625 When this is called by a <code class="command">COPY FROM</code> command, the626 plan-related global data in <code class="literal">mtstate</code> is not provided627 and the <code class="literal">planSlot</code> parameter of628 <code class="function">ExecForeignInsert</code> subsequently called for each629 inserted tuple is <code class="literal">NULL</code>, whether the foreign table is630 the partition chosen for tuple routing or the target specified in the631 command.632 </p><p>633 If the <code class="function">BeginForeignInsert</code> pointer is set to634 <code class="literal">NULL</code>, no action is taken for the initialization.635 </p><p>636 Note that if the FDW does not support routable foreign-table partitions637 and/or executing <code class="command">COPY FROM</code> on foreign tables, this638 function or <code class="function">ExecForeignInsert/ExecForeignBatchInsert</code>639 subsequently called must throw error as needed.640 </p><p>641</p><pre class="programlisting">642void643EndForeignInsert(EState *estate,644 ResultRelInfo *rinfo);645</pre><p>646 647 End the insert operation and release resources. It is normally not important648 to release palloc'd memory, but for example open files and connections649 to remote servers should be cleaned up.650 </p><p>651 If the <code class="function">EndForeignInsert</code> pointer is set to652 <code class="literal">NULL</code>, no action is taken for the termination.653 </p><p>654</p><pre class="programlisting">655int656IsForeignRelUpdatable(Relation rel);657</pre><p>658 659 Report which update operations the specified foreign table supports.660 The return value should be a bit mask of rule event numbers indicating661 which operations are supported by the foreign table, using the662 <code class="literal">CmdType</code> enumeration; that is,663 <code class="literal">(1 << CMD_UPDATE) = 4</code> for <code class="command">UPDATE</code>,664 <code class="literal">(1 << CMD_INSERT) = 8</code> for <code class="command">INSERT</code>, and665 <code class="literal">(1 << CMD_DELETE) = 16</code> for <code class="command">DELETE</code>.666 </p><p>667 If the <code class="function">IsForeignRelUpdatable</code> pointer is set to668 <code class="literal">NULL</code>, foreign tables are assumed to be insertable, updatable,669 or deletable if the FDW provides <code class="function">ExecForeignInsert</code>,670 <code class="function">ExecForeignUpdate</code>, or <code class="function">ExecForeignDelete</code>671 respectively. This function is only needed if the FDW supports some672 tables that are updatable and some that are not. (Even then, it's673 permissible to throw an error in the execution routine instead of674 checking in this function. However, this function is used to determine675 updatability for display in the <code class="literal">information_schema</code> views.)676 </p><p>677 Some inserts, updates, and deletes to foreign tables can be optimized678 by implementing an alternative set of interfaces. The ordinary679 interfaces for inserts, updates, and deletes fetch rows from the remote680 server and then modify those rows one at a time. In some cases, this681 row-by-row approach is necessary, but it can be inefficient. If it is682 possible for the foreign server to determine which rows should be683 modified without actually retrieving them, and if there are no local684 structures which would affect the operation (row-level local triggers,685 stored generated columns, or <code class="literal">WITH CHECK OPTION</code>686 constraints from parent views), then it is possible to arrange things687 so that the entire operation is performed on the remote server. The688 interfaces described below make this possible.689 </p><p>690</p><pre class="programlisting">691bool692PlanDirectModify(PlannerInfo *root,693 ModifyTable *plan,694 Index resultRelation,695 int subplan_index);696</pre><p>697 698 Decide whether it is safe to execute a direct modification699 on the remote server. If so, return <code class="literal">true</code> after performing700 planning actions needed for that. Otherwise, return <code class="literal">false</code>.701 This optional function is called during query planning.702 If this function succeeds, <code class="function">BeginDirectModify</code>,703 <code class="function">IterateDirectModify</code> and <code class="function">EndDirectModify</code> will704 be called at the execution stage, instead. Otherwise, the table705 modification will be executed using the table-updating functions706 described above.707 The parameters are the same as for <code class="function">PlanForeignModify</code>.708 </p><p>709 To execute the direct modification on the remote server, this function710 must rewrite the target subplan with a <code class="structname">ForeignScan</code> plan711 node that executes the direct modification on the remote server. The712 <code class="structfield">operation</code> and <code class="structfield">resultRelation</code> fields713 of the <code class="structname">ForeignScan</code> must be set appropriately.714 <code class="structfield">operation</code> must be set to the <code class="literal">CmdType</code>715 enumeration corresponding to the statement kind (that is,716 <code class="literal">CMD_UPDATE</code> for <code class="command">UPDATE</code>,717 <code class="literal">CMD_INSERT</code> for <code class="command">INSERT</code>, and718 <code class="literal">CMD_DELETE</code> for <code class="command">DELETE</code>), and the719 <code class="literal">resultRelation</code> argument must be copied to the720 <code class="structfield">resultRelation</code> field.721 </p><p>722 See <a class="xref" href="fdw-planning.html" title="59.4. Foreign Data Wrapper Query Planning">Section 59.4</a> for additional information.723 </p><p>724 If the <code class="function">PlanDirectModify</code> pointer is set to725 <code class="literal">NULL</code>, no attempts to execute a direct modification on the726 remote server are taken.727 </p><p>728</p><pre class="programlisting">729void730BeginDirectModify(ForeignScanState *node,731 int eflags);732</pre><p>733 734 Prepare to execute a direct modification on the remote server.735 This is called during executor startup. It should perform any736 initialization needed prior to the direct modification (that should be737 done upon the first call to <code class="function">IterateDirectModify</code>).738 The <code class="structname">ForeignScanState</code> node has already been created, but739 its <code class="structfield">fdw_state</code> field is still NULL. Information about740 the table to modify is accessible through the741 <code class="structname">ForeignScanState</code> node (in particular, from the underlying742 <code class="structname">ForeignScan</code> plan node, which contains any FDW-private743 information provided by <code class="function">PlanDirectModify</code>).744 <code class="literal">eflags</code> contains flag bits describing the executor's745 operating mode for this plan node.746 </p><p>747 Note that when <code class="literal">(eflags & EXEC_FLAG_EXPLAIN_ONLY)</code> is748 true, this function should not perform any externally-visible actions;749 it should only do the minimum required to make the node state valid750 for <code class="function">ExplainDirectModify</code> and <code class="function">EndDirectModify</code>.751 </p><p>752 If the <code class="function">BeginDirectModify</code> pointer is set to753 <code class="literal">NULL</code>, no attempts to execute a direct modification on the754 remote server are taken.755 </p><p>756</p><pre class="programlisting">757TupleTableSlot *758IterateDirectModify(ForeignScanState *node);759</pre><p>760 761 When the <code class="command">INSERT</code>, <code class="command">UPDATE</code> or <code class="command">DELETE</code>762 query doesn't have a <code class="literal">RETURNING</code> clause, just return NULL763 after a direct modification on the remote server.764 When the query has the clause, fetch one result containing the data765 needed for the <code class="literal">RETURNING</code> calculation, returning it in a766 tuple table slot (the node's <code class="structfield">ScanTupleSlot</code> should be767 used for this purpose). The data that was actually inserted, updated768 or deleted must be stored in769 <code class="literal">node->resultRelInfo->ri_projectReturning->pi_exprContext->ecxt_scantuple</code>.770 Return NULL if no more rows are available.771 Note that this is called in a short-lived memory context that will be772 reset between invocations. Create a memory context in773 <code class="function">BeginDirectModify</code> if you need longer-lived storage, or use774 the <code class="structfield">es_query_cxt</code> of the node's <code class="structname">EState</code>.775 </p><p>776 The rows returned must match the <code class="structfield">fdw_scan_tlist</code> target777 list if one was supplied, otherwise they must match the row type of the778 foreign table being updated. If you choose to optimize away fetching779 columns that are not needed for the <code class="literal">RETURNING</code> calculation,780 you should insert nulls in those column positions, or else generate a781 <code class="structfield">fdw_scan_tlist</code> list with those columns omitted.782 </p><p>783 Whether the query has the clause or not, the query's reported row count784 must be incremented by the FDW itself. When the query doesn't have the785 clause, the FDW must also increment the row count for the786 <code class="structname">ForeignScanState</code> node in the <code class="command">EXPLAIN ANALYZE</code>787 case.788 </p><p>789 If the <code class="function">IterateDirectModify</code> pointer is set to790 <code class="literal">NULL</code>, no attempts to execute a direct modification on the791 remote server are taken.792 </p><p>793</p><pre class="programlisting">794void795EndDirectModify(ForeignScanState *node);796</pre><p>797 798 Clean up following a direct modification on the remote server. It is799 normally not important to release palloc'd memory, but for example open800 files and connections to the remote server should be cleaned up.801 </p><p>802 If the <code class="function">EndDirectModify</code> pointer is set to803 <code class="literal">NULL</code>, no attempts to execute a direct modification on the804 remote server are taken.805 </p></div><div class="sect2" id="FDW-CALLBACKS-TRUNCATE"><div class="titlepage"><div><div><h3 class="title">59.2.5. FDW Routines for <code class="command">TRUNCATE</code> <a href="#FDW-CALLBACKS-TRUNCATE" class="id_link">#</a></h3></div></div></div><p>806</p><pre class="programlisting">807void808ExecForeignTruncate(List *rels,809 DropBehavior behavior,810 bool restart_seqs);811</pre><p>812 813 Truncate foreign tables. This function is called when814 <a class="xref" href="sql-truncate.html" title="TRUNCATE"><span class="refentrytitle">TRUNCATE</span></a> is executed on a foreign table.815 <code class="literal">rels</code> is a list of <code class="structname">Relation</code>816 data structures of foreign tables to truncate.817 </p><p>818 <code class="literal">behavior</code> is either <code class="literal">DROP_RESTRICT</code>819 or <code class="literal">DROP_CASCADE</code> indicating that the820 <code class="literal">RESTRICT</code> or <code class="literal">CASCADE</code> option was821 requested in the original <code class="command">TRUNCATE</code> command,822 respectively.823 </p><p>824 If <code class="literal">restart_seqs</code> is <code class="literal">true</code>,825 the original <code class="command">TRUNCATE</code> command requested the826 <code class="literal">RESTART IDENTITY</code> behavior, otherwise the827 <code class="literal">CONTINUE IDENTITY</code> behavior was requested.828 </p><p>829 Note that the <code class="literal">ONLY</code> options specified830 in the original <code class="command">TRUNCATE</code> command are not passed to831 <code class="function">ExecForeignTruncate</code>. This behavior is similar to832 the callback functions of <code class="command">SELECT</code>,833 <code class="command">UPDATE</code> and <code class="command">DELETE</code> on834 a foreign table.835 </p><p>836 <code class="function">ExecForeignTruncate</code> is invoked once per837 foreign server for which foreign tables are to be truncated.838 This means that all foreign tables included in <code class="literal">rels</code>839 must belong to the same server.840 </p><p>841 If the <code class="function">ExecForeignTruncate</code> pointer is set to842 <code class="literal">NULL</code>, attempts to truncate foreign tables will843 fail with an error message.844 </p></div><div class="sect2" id="FDW-CALLBACKS-ROW-LOCKING"><div class="titlepage"><div><div><h3 class="title">59.2.6. FDW Routines for Row Locking <a href="#FDW-CALLBACKS-ROW-LOCKING" class="id_link">#</a></h3></div></div></div><p>845 If an FDW wishes to support <em class="firstterm">late row locking</em> (as described846 in <a class="xref" href="fdw-row-locking.html" title="59.5. Row Locking in Foreign Data Wrappers">Section 59.5</a>), it must provide the following847 callback functions:848 </p><p>849</p><pre class="programlisting">850RowMarkType851GetForeignRowMarkType(RangeTblEntry *rte,852 LockClauseStrength strength);853</pre><p>854 855 Report which row-marking option to use for a foreign table.856 <code class="literal">rte</code> is the <code class="structname">RangeTblEntry</code> node for the table857 and <code class="literal">strength</code> describes the lock strength requested by the858 relevant <code class="literal">FOR UPDATE/SHARE</code> clause, if any. The result must be859 a member of the <code class="literal">RowMarkType</code> enum type.860 </p><p>861 This function is called during query planning for each foreign table that862 appears in an <code class="command">UPDATE</code>, <code class="command">DELETE</code>, or <code class="command">SELECT863 FOR UPDATE/SHARE</code> query and is not the target of <code class="command">UPDATE</code>864 or <code class="command">DELETE</code>.865 </p><p>866 If the <code class="function">GetForeignRowMarkType</code> pointer is set to867 <code class="literal">NULL</code>, the <code class="literal">ROW_MARK_COPY</code> option is always used.868 (This implies that <code class="function">RefetchForeignRow</code> will never be called,869 so it need not be provided either.)870 </p><p>871 See <a class="xref" href="fdw-row-locking.html" title="59.5. Row Locking in Foreign Data Wrappers">Section 59.5</a> for more information.872 </p><p>873</p><pre class="programlisting">874void875RefetchForeignRow(EState *estate,876 ExecRowMark *erm,877 Datum rowid,878 TupleTableSlot *slot,879 bool *updated);880</pre><p>881 882 Re-fetch one tuple slot from the foreign table, after locking it if required.883 <code class="literal">estate</code> is global execution state for the query.884 <code class="literal">erm</code> is the <code class="structname">ExecRowMark</code> struct describing885 the target foreign table and the row lock type (if any) to acquire.886 <code class="literal">rowid</code> identifies the tuple to be fetched.887 <code class="literal">slot</code> contains nothing useful upon call, but can be used to888 hold the returned tuple. <code class="literal">updated</code> is an output parameter.889 </p><p>890 This function should store the tuple into the provided slot, or clear it if891 the row lock couldn't be obtained. The row lock type to acquire is892 defined by <code class="literal">erm->markType</code>, which is the value893 previously returned by <code class="function">GetForeignRowMarkType</code>.894 (<code class="literal">ROW_MARK_REFERENCE</code> means to just re-fetch the tuple895 without acquiring any lock, and <code class="literal">ROW_MARK_COPY</code> will896 never be seen by this routine.)897 </p><p>898 In addition, <code class="literal">*updated</code> should be set to <code class="literal">true</code>899 if what was fetched was an updated version of the tuple rather than900 the same version previously obtained. (If the FDW cannot be sure about901 this, always returning <code class="literal">true</code> is recommended.)902 </p><p>903 Note that by default, failure to acquire a row lock should result in904 raising an error; returning with an empty slot is only appropriate if905 the <code class="literal">SKIP LOCKED</code> option is specified906 by <code class="literal">erm->waitPolicy</code>.907 </p><p>908 The <code class="literal">rowid</code> is the <code class="structfield">ctid</code> value previously read909 for the row to be re-fetched. Although the <code class="literal">rowid</code> value is910 passed as a <code class="type">Datum</code>, it can currently only be a <code class="type">tid</code>. The911 function API is chosen in hopes that it may be possible to allow other912 data types for row IDs in future.913 </p><p>914 If the <code class="function">RefetchForeignRow</code> pointer is set to915 <code class="literal">NULL</code>, attempts to re-fetch rows will fail916 with an error message.917 </p><p>918 See <a class="xref" href="fdw-row-locking.html" title="59.5. Row Locking in Foreign Data Wrappers">Section 59.5</a> for more information.919 </p><p>920</p><pre class="programlisting">921bool922RecheckForeignScan(ForeignScanState *node,923 TupleTableSlot *slot);924</pre><p>925 Recheck that a previously-returned tuple still matches the relevant926 scan and join qualifiers, and possibly provide a modified version of927 the tuple. For foreign data wrappers which do not perform join pushdown,928 it will typically be more convenient to set this to <code class="literal">NULL</code> and929 instead set <code class="structfield">fdw_recheck_quals</code> appropriately.930 When outer joins are pushed down, however, it isn't sufficient to931 reapply the checks relevant to all the base tables to the result tuple,932 even if all needed attributes are present, because failure to match some933 qualifier might result in some attributes going to NULL, rather than in934 no tuple being returned. <code class="literal">RecheckForeignScan</code> can recheck935 qualifiers and return true if they are still satisfied and false936 otherwise, but it can also store a replacement tuple into the supplied937 slot.938 </p><p>939 To implement join pushdown, a foreign data wrapper will typically940 construct an alternative local join plan which is used only for941 rechecks; this will become the outer subplan of the942 <code class="literal">ForeignScan</code>. When a recheck is required, this subplan943 can be executed and the resulting tuple can be stored in the slot.944 This plan need not be efficient since no base table will return more945 than one row; for example, it may implement all joins as nested loops.946 The function <code class="literal">GetExistingLocalJoinPath</code> may be used to search947 existing paths for a suitable local join path, which can be used as the948 alternative local join plan. <code class="literal">GetExistingLocalJoinPath</code>949 searches for an unparameterized path in the path list of the specified950 join relation. (If it does not find such a path, it returns NULL, in951 which case a foreign data wrapper may build the local path by itself or952 may choose not to create access paths for that join.)953 </p></div><div class="sect2" id="FDW-CALLBACKS-EXPLAIN"><div class="titlepage"><div><div><h3 class="title">59.2.7. FDW Routines for <code class="command">EXPLAIN</code> <a href="#FDW-CALLBACKS-EXPLAIN" class="id_link">#</a></h3></div></div></div><p>954</p><pre class="programlisting">955void956ExplainForeignScan(ForeignScanState *node,957 ExplainState *es);958</pre><p>959 960 Print additional <code class="command">EXPLAIN</code> output for a foreign table scan.961 This function can call <code class="function">ExplainPropertyText</code> and962 related functions to add fields to the <code class="command">EXPLAIN</code> output.963 The flag fields in <code class="literal">es</code> can be used to determine what to964 print, and the state of the <code class="structname">ForeignScanState</code> node965 can be inspected to provide run-time statistics in the <code class="command">EXPLAIN966 ANALYZE</code> case.967 </p><p>968 If the <code class="function">ExplainForeignScan</code> pointer is set to969 <code class="literal">NULL</code>, no additional information is printed during970 <code class="command">EXPLAIN</code>.971 </p><p>972</p><pre class="programlisting">973void974ExplainForeignModify(ModifyTableState *mtstate,975 ResultRelInfo *rinfo,976 List *fdw_private,977 int subplan_index,978 struct ExplainState *es);979</pre><p>980 981 Print additional <code class="command">EXPLAIN</code> output for a foreign table update.982 This function can call <code class="function">ExplainPropertyText</code> and983 related functions to add fields to the <code class="command">EXPLAIN</code> output.984 The flag fields in <code class="literal">es</code> can be used to determine what to985 print, and the state of the <code class="structname">ModifyTableState</code> node986 can be inspected to provide run-time statistics in the <code class="command">EXPLAIN987 ANALYZE</code> case. The first four arguments are the same as for988 <code class="function">BeginForeignModify</code>.989 </p><p>990 If the <code class="function">ExplainForeignModify</code> pointer is set to991 <code class="literal">NULL</code>, no additional information is printed during992 <code class="command">EXPLAIN</code>.993 </p><p>994</p><pre class="programlisting">995void996ExplainDirectModify(ForeignScanState *node,997 ExplainState *es);998</pre><p>999 1000 Print additional <code class="command">EXPLAIN</code> output for a direct modification1001 on the remote server.1002 This function can call <code class="function">ExplainPropertyText</code> and1003 related functions to add fields to the <code class="command">EXPLAIN</code> output.1004 The flag fields in <code class="literal">es</code> can be used to determine what to1005 print, and the state of the <code class="structname">ForeignScanState</code> node1006 can be inspected to provide run-time statistics in the <code class="command">EXPLAIN1007 ANALYZE</code> case.1008 </p><p>1009 If the <code class="function">ExplainDirectModify</code> pointer is set to1010 <code class="literal">NULL</code>, no additional information is printed during1011 <code class="command">EXPLAIN</code>.1012 </p></div><div class="sect2" id="FDW-CALLBACKS-ANALYZE"><div class="titlepage"><div><div><h3 class="title">59.2.8. FDW Routines for <code class="command">ANALYZE</code> <a href="#FDW-CALLBACKS-ANALYZE" class="id_link">#</a></h3></div></div></div><p>1013</p><pre class="programlisting">1014bool1015AnalyzeForeignTable(Relation relation,1016 AcquireSampleRowsFunc *func,1017 BlockNumber *totalpages);1018</pre><p>1019 1020 This function is called when <a class="xref" href="sql-analyze.html" title="ANALYZE"><span class="refentrytitle">ANALYZE</span></a> is executed on1021 a foreign table. If the FDW can collect statistics for this1022 foreign table, it should return <code class="literal">true</code>, and provide a pointer1023 to a function that will collect sample rows from the table in1024 <em class="parameter"><code>func</code></em>, plus the estimated size of the table in pages in1025 <em class="parameter"><code>totalpages</code></em>. Otherwise, return <code class="literal">false</code>.1026 </p><p>1027 If the FDW does not support collecting statistics for any tables, the1028 <code class="function">AnalyzeForeignTable</code> pointer can be set to <code class="literal">NULL</code>.1029 </p><p>1030 If provided, the sample collection function must have the signature1031</p><pre class="programlisting">1032int1033AcquireSampleRowsFunc(Relation relation,1034 int elevel,1035 HeapTuple *rows,1036 int targrows,1037 double *totalrows,1038 double *totaldeadrows);1039</pre><p>1040 1041 A random sample of up to <em class="parameter"><code>targrows</code></em> rows should be collected1042 from the table and stored into the caller-provided <em class="parameter"><code>rows</code></em>1043 array. The actual number of rows collected must be returned. In1044 addition, store estimates of the total numbers of live and dead rows in1045 the table into the output parameters <em class="parameter"><code>totalrows</code></em> and1046 <em class="parameter"><code>totaldeadrows</code></em>. (Set <em class="parameter"><code>totaldeadrows</code></em> to zero1047 if the FDW does not have any concept of dead rows.)1048 </p></div><div class="sect2" id="FDW-CALLBACKS-IMPORT"><div class="titlepage"><div><div><h3 class="title">59.2.9. FDW Routines for <code class="command">IMPORT FOREIGN SCHEMA</code> <a href="#FDW-CALLBACKS-IMPORT" class="id_link">#</a></h3></div></div></div><p>1049</p><pre class="programlisting">1050List *1051ImportForeignSchema(ImportForeignSchemaStmt *stmt, Oid serverOid);1052</pre><p>1053 1054 Obtain a list of foreign table creation commands. This function is1055 called when executing <a class="xref" href="sql-importforeignschema.html" title="IMPORT FOREIGN SCHEMA"><span class="refentrytitle">IMPORT FOREIGN SCHEMA</span></a>, and is1056 passed the parse tree for that statement, as well as the OID of the1057 foreign server to use. It should return a list of C strings, each of1058 which must contain a <a class="xref" href="sql-createforeigntable.html" title="CREATE FOREIGN TABLE"><span class="refentrytitle">CREATE FOREIGN TABLE</span></a> command.1059 These strings will be parsed and executed by the core server.1060 </p><p>1061 Within the <code class="structname">ImportForeignSchemaStmt</code> struct,1062 <code class="structfield">remote_schema</code> is the name of the remote schema from1063 which tables are to be imported.1064 <code class="structfield">list_type</code> identifies how to filter table names:1065 <code class="literal">FDW_IMPORT_SCHEMA_ALL</code> means that all tables in the remote1066 schema should be imported (in this case <code class="structfield">table_list</code> is1067 empty), <code class="literal">FDW_IMPORT_SCHEMA_LIMIT_TO</code> means to include only1068 tables listed in <code class="structfield">table_list</code>,1069 and <code class="literal">FDW_IMPORT_SCHEMA_EXCEPT</code> means to exclude the tables1070 listed in <code class="structfield">table_list</code>.1071 <code class="structfield">options</code> is a list of options used for the import process.1072 The meanings of the options are up to the FDW.1073 For example, an FDW could use an option to define whether the1074 <code class="literal">NOT NULL</code> attributes of columns should be imported.1075 These options need not have anything to do with those supported by the1076 FDW as database object options.1077 </p><p>1078 The FDW may ignore the <code class="structfield">local_schema</code> field of1079 the <code class="structname">ImportForeignSchemaStmt</code>, because the core server1080 will automatically insert that name into the parsed <code class="command">CREATE1081 FOREIGN TABLE</code> commands.1082 </p><p>1083 The FDW does not have to concern itself with implementing the filtering1084 specified by <code class="structfield">list_type</code> and <code class="structfield">table_list</code>,1085 either, as the core server will automatically skip any returned commands1086 for tables excluded according to those options. However, it's often1087 useful to avoid the work of creating commands for excluded tables in the1088 first place. The function <code class="function">IsImportableForeignTable()</code> may be1089 useful to test whether a given foreign-table name will pass the filter.1090 </p><p>1091 If the FDW does not support importing table definitions, the1092 <code class="function">ImportForeignSchema</code> pointer can be set to <code class="literal">NULL</code>.1093 </p></div><div class="sect2" id="FDW-CALLBACKS-PARALLEL"><div class="titlepage"><div><div><h3 class="title">59.2.10. FDW Routines for Parallel Execution <a href="#FDW-CALLBACKS-PARALLEL" class="id_link">#</a></h3></div></div></div><p>1094 A <code class="structname">ForeignScan</code> node can, optionally, support parallel1095 execution. A parallel <code class="structname">ForeignScan</code> will be executed1096 in multiple processes and must return each row exactly once across1097 all cooperating processes. To do this, processes can coordinate through1098 fixed-size chunks of dynamic shared memory. This shared memory is not1099 guaranteed to be mapped at the same address in every process, so it1100 must not contain pointers. The following functions are all optional,1101 but most are required if parallel execution is to be supported.1102 </p><p>1103</p><pre class="programlisting">1104bool1105IsForeignScanParallelSafe(PlannerInfo *root, RelOptInfo *rel,1106 RangeTblEntry *rte);1107</pre><p>1108 Test whether a scan can be performed within a parallel worker. This1109 function will only be called when the planner believes that a parallel1110 plan might be possible, and should return true if it is safe for that scan1111 to run within a parallel worker. This will generally not be the case if1112 the remote data source has transaction semantics, unless the worker's1113 connection to the data can somehow be made to share the same transaction1114 context as the leader.1115 </p><p>1116 If this function is not defined, it is assumed that the scan must take1117 place within the parallel leader. Note that returning true does not mean1118 that the scan itself can be done in parallel, only that the scan can be1119 performed within a parallel worker. Therefore, it can be useful to define1120 this method even when parallel execution is not supported.1121 </p><p>1122</p><pre class="programlisting">1123Size1124EstimateDSMForeignScan(ForeignScanState *node, ParallelContext *pcxt);1125</pre><p>1126 Estimate the amount of dynamic shared memory that will be required1127 for parallel operation. This may be higher than the amount that will1128 actually be used, but it must not be lower. The return value is in bytes.1129 This function is optional, and can be omitted if not needed; but if it1130 is omitted, the next three functions must be omitted as well, because1131 no shared memory will be allocated for the FDW's use.1132 </p><p>1133</p><pre class="programlisting">1134void1135InitializeDSMForeignScan(ForeignScanState *node, ParallelContext *pcxt,1136 void *coordinate);1137</pre><p>1138 Initialize the dynamic shared memory that will be required for parallel1139 operation. <code class="literal">coordinate</code> points to a shared memory area of1140 size equal to the return value of <code class="function">EstimateDSMForeignScan</code>.1141 This function is optional, and can be omitted if not needed.1142 </p><p>1143</p><pre class="programlisting">1144void1145ReInitializeDSMForeignScan(ForeignScanState *node, ParallelContext *pcxt,1146 void *coordinate);1147</pre><p>1148 Re-initialize the dynamic shared memory required for parallel operation1149 when the foreign-scan plan node is about to be re-scanned.1150 This function is optional, and can be omitted if not needed.1151 Recommended practice is that this function reset only shared state,1152 while the <code class="function">ReScanForeignScan</code> function resets only local1153 state. Currently, this function will be called1154 before <code class="function">ReScanForeignScan</code>, but it's best not to rely on1155 that ordering.1156 </p><p>1157</p><pre class="programlisting">1158void1159InitializeWorkerForeignScan(ForeignScanState *node, shm_toc *toc,1160 void *coordinate);1161</pre><p>1162 Initialize a parallel worker's local state based on the shared state1163 set up by the leader during <code class="function">InitializeDSMForeignScan</code>.1164 This function is optional, and can be omitted if not needed.1165 </p><p>1166</p><pre class="programlisting">1167void1168ShutdownForeignScan(ForeignScanState *node);1169</pre><p>1170 Release resources when it is anticipated the node will not be executed1171 to completion. This is not called in all cases; sometimes,1172 <code class="literal">EndForeignScan</code> may be called without this function having1173 been called first. Since the DSM segment used by parallel query is1174 destroyed just after this callback is invoked, foreign data wrappers that1175 wish to take some action before the DSM segment goes away should implement1176 this method.1177 </p></div><div class="sect2" id="FDW-CALLBACKS-ASYNC"><div class="titlepage"><div><div><h3 class="title">59.2.11. FDW Routines for Asynchronous Execution <a href="#FDW-CALLBACKS-ASYNC" class="id_link">#</a></h3></div></div></div><p>1178 A <code class="structname">ForeignScan</code> node can, optionally, support1179 asynchronous execution as described in1180 <code class="filename">src/backend/executor/README</code>. The following1181 functions are all optional, but are all required if asynchronous1182 execution is to be supported.1183 </p><p>1184</p><pre class="programlisting">1185bool1186IsForeignPathAsyncCapable(ForeignPath *path);1187</pre><p>1188 Test whether a given <code class="structname">ForeignPath</code> path can scan1189 the underlying foreign relation asynchronously.1190 This function will only be called at the end of query planning when the1191 given path is a direct child of an <code class="structname">AppendPath</code>1192 path and when the planner believes that asynchronous execution improves1193 performance, and should return true if the given path is able to scan the1194 foreign relation asynchronously.1195 </p><p>1196 If this function is not defined, it is assumed that the given path scans1197 the foreign relation using <code class="function">IterateForeignScan</code>.1198 (This implies that the callback functions described below will never be1199 called, so they need not be provided either.)1200 </p><p>