Complete Linux

home *** CD-ROM | disk | FTP | other *** search

/ Complete Linux / Complete Linux.iso / docs / apps / database / postgres / postgre4.z / postgre4 / src / executor / ex_main.c < prev next >

Wrap

C/C++ Source or Header | 1992-08-27 | 36.6 KB | 1,360 lines

/* ---------------------------------------------------------------- * FILE * ex_main.c * * DESCRIPTION * top level executor interface routines * * INTERFACE ROUTINES * ExecMain - main interface to executor * * NOTES * * IDENTIFICATION * $Header: /private/postgres/src/executor/RCS/ex_main.c,v 1.23 1992/08/26 23:18:06 mer Exp $ * ---------------------------------------------------------------- */ #define EXEC_RULEMANAGER 1 #include "tcop/slaves.h" #include "executor/executor.h" RcsId("$Header: /private/postgres/src/executor/RCS/ex_main.c,v 1.23 1992/08/26 23:18:06 mer Exp $"); /* ---------------------------------------------------------------- * random junk * ---------------------------------------------------------------- */ bool ExecIsInitialized = false; Const ConstTrue; Const ConstFalse; extern int Quiet; extern List rule; /* ---------------------------------------------------------------- * InitializeExecutor * * This is only called once - by ExecMain, the first time * the user executes a query. * ---------------------------------------------------------------- */ void InitializeExecutor() { MemoryContext oldContext; ObjectId boolType = 16; /* see local.bki */ Size boolLen = 1; /* these should come from lookups */ Datum datumTrue; Datum datumFalse; /* ---------------- * make certain we are in the top memory context * or else this stuff will go away after the first transaction. * ---------------- */ oldContext = MemoryContextSwitchTo(TopMemoryContext); /* ---------------- * initialize ConstTrue and ConstFalse * ---------------- */ datumTrue = Int32GetDatum((int32) true); datumFalse = Int32GetDatum((int32) false); ConstTrue = MakeConst(boolType, boolLen, datumTrue, false, true); ConstFalse = MakeConst(boolType, boolLen, datumFalse, false, true); #ifdef EXEC_DEBUGVARIABLEFILE /* ---------------- * initialize debugging variables * ---------------- */ if (! Quiet) printf("==== using %s\n", EXEC_DEBUGVARIABLEFILE); DebugVariableFileSet(EXEC_DEBUGVARIABLEFILE); #endif EXEC_DEBUGVARIABLEFILE /* ---------------- * return to old memory context * ---------------- */ (void) MemoryContextSwitchTo(oldContext); ExecIsInitialized = true; } /* ---------------------------------------------------------------- * ExecMain * * This is the main routine of the executor module. It accepts * the query descriptor from the traffic cop and executes the * query plan. * * Parameter: * argList --(commandType parseTree queryPlan queryState queryFeature) * ---------------------------------------------------------------- */ List ExecMain(queryDesc, estate, feature) List queryDesc; EState estate; List feature; { int operation; List parseTree; Plan plan; List result; CommandDest dest; void (*destination)(); /* ---------------- * sanity checks * ---------------- */ Assert(!lispNullp(queryDesc)); /* ---------------- * initialize if necessary * ---------------- */ if (!ExecIsInitialized) InitializeExecutor(); /* ---------------- * extract information from the query descriptor * and the query feature. * ---------------- */ operation = CAtom(GetOperation(queryDesc)); parseTree = QdGetParseTree(queryDesc); plan = QdGetPlan(queryDesc); dest = QdGetDest(queryDesc); destination = (void (*)()) DestToFunction(dest); /* ---------------- * ExecMain is typically called from the tcop three * times: * * first time, command = EXEC_START initialize * second time, command = EXEC_RUN run the plan * last time, command = EXEC_END cleanup * ---------------- */ result = LispNil; switch(CInteger(FeatureGetCommand(feature))) { /* ---------------- * EXEC_START, EXEC_RUN, EXEC_END are the * most general cases.. * ---------------- */ case EXEC_START: /* result is a tuple descriptor */ result = InitPlan(operation, parseTree, plan, estate); /* reset buffer refcount. the current refcounts * are saved and will be restored at the end of * this ExecMain cycle. * * this makes sure that when ExecMain's are * called recursively as for tuple level rules * or postquel functions, the buffers pinned * by one ExecMain will not be unpinned by another * ExecMain. */ BufferRefCountReset(get_es_refcount(estate)); break; case EXEC_RUN: /* result is a tuple or LispNil */ result = (List) ExecutePlan(estate, plan, parseTree, operation, ALL_TUPLES, EXEC_FRWD, destination); break; case EXEC_END: /* result is a tuple or LispNil */ result = EndPlan(plan, estate); /* restore saved refcounts. */ BufferRefCountRestore(get_es_refcount(estate)); break; /* ---------------- * retrieve next n "forward" tuples * ---------------- */ case EXEC_FOR: /* result is a tuple or LispNil */ result = (List) ExecutePlan(estate, plan, parseTree, operation, CInteger(FeatureGetCount(feature)), EXEC_FRWD, destination); break; /* ---------------- * retrieve next n "backward" tuples * ---------------- */ case EXEC_BACK: /* result is a tuple or LispNil */ result = (List) ExecutePlan(estate, plan, parseTree, operation, CInteger(FeatureGetCount(feature)), EXEC_BKWD, destination); break; /* ---------------- * return one tuple but don't "retrieve" it. * (this is used by the rule manager..) -cim 9/14/89 * ---------------- */ case EXEC_RETONE: /* result is a tuple or LispNil */ result = (List) ExecutePlan(estate, plan, parseTree, operation, ONE_TUPLE, EXEC_FRWD, destination); break; default: elog(DEBUG, "ExecMain: Unknown command."); break; } return result; } /* ---------------------------------------------------------------- * InitPlan * * Initialises the query plan: open files, allocate storage * and start up the rule manager * ---------------------------------------------------------------- */ extern Relation CopyRelDescUsing(); List InitPlan(operation, parseTree, plan, estate) int operation; EState estate; List parseTree; Plan plan; { List parseRoot; List rangeTable; List resultRelation; Relation relationRelationDesc; Relation intoRelationDesc; TupleDescriptor tupType; List targetList; int len; List attinfo; extern List MakeList(); #ifdef PALLOC_DEBUG { extern int query_tuple_count; query_tuple_count = 0; } #endif PALLOC_DEBUG /* ---------------- * get information from query descriptor * ---------------- */ parseRoot = parse_root(parseTree); rangeTable = parse_tree_range_table(parseTree); resultRelation = parse_tree_result_relation(parseTree); /* ---------------- * initialize the node's execution state * ---------------- */ set_es_range_table(estate, rangeTable); set_es_error_message(estate, (Name) "Foo bar"); /* ---------------- * initialize the BaseId counter so node base_id's * are assigned correctly. Someday baseid's will have to * be stored someplace other than estate because they * should be unique per query planned. * ---------------- */ set_es_BaseId(estate, 1); /* ---------------- * initialize result relation stuff * ---------------- */ set_es_result_rel_scanstate(estate, NULL); if (resultRelation != NULL && operation != RETRIEVE) { /* ---------------- * if we have a result relation, open the it and * initialize the result relation info stuff. * ---------------- */ RelationInfo resultRelationInfo; Index resultRelationIndex; List rtentry; ObjectId resultRelationOid; Relation resultRelationDesc; RelationRuleInfo resRelRuleInfo; resultRelationIndex = CInteger(resultRelation); rtentry = rt_fetch(resultRelationIndex, rangeTable); resultRelationOid = CInteger(rt_relid(rtentry)); resultRelationDesc = heap_open(resultRelationOid); resultRelationInfo = MakeRelationInfo(resultRelationIndex, resultRelationDesc, 0, /* num indices */ NULL, /* index descs */ NULL); /* index key info */ /* ---------------- * open indices on result relation and save descriptors * in the result relation information.. * ---------------- */ ExecOpenIndices(resultRelationOid, resultRelationInfo); set_es_result_relation_info(estate, resultRelationInfo); /* ---------------- * Initialize the result relation rule info. * ---------------- */ resRelRuleInfo = prs2MakeRelationRuleInfo(resultRelationDesc, operation); set_es_result_rel_ruleinfo(estate, resRelRuleInfo); } else { /* ---------------- * if no result relation, then set state appropriately * ---------------- */ set_es_result_relation_info(estate, NULL); } relationRelationDesc = heap_openr(RelationRelationName); set_es_relation_relation_descriptor(estate, relationRelationDesc); /* ---------------- * initialize the executor "tuple" table. * ---------------- */ { int nSlots = ExecCountSlotsNode(plan); TupleTable tupleTable = ExecCreateTupleTable(nSlots+10); set_es_tupleTable(estate, tupleTable); } /* ---------------- * initialize the private state information for * all the nodes in the query tree. This opens * files, allocates storage and leaves us ready * to start processing tuples.. * ---------------- */ ExecInitNode(plan, estate, NULL); /* ---------------- * get the tuple descriptor describing the type * of tuples to return.. (this is especially important * if we are creating a relation with "retrieve into") * ---------------- */ tupType = ExecGetTupType(plan); /* tuple descriptor */ targetList = get_qptargetlist(plan); len = ExecTargetListLength(targetList); /* number of attributes */ /* ---------------- * now that we have the target list, initialize the junk filter * if this is a REPLACE or a DELETE query. * We also init the junk filter if this is an append query * (there might be some rule lock info there...) * NOTE: in the future we might want to initialize the junk * filter for all queries. * ---------------- */ if (operation == REPLACE || operation == DELETE || operation == APPEND) { JunkFilter j = (JunkFilter) ExecInitJunkFilter(targetList); set_es_junkFilter(estate, j); } else set_es_junkFilter(estate, NULL); /* ---------------- * initialize the "into" relation * ---------------- */ intoRelationDesc = (Relation) NULL; if (operation == RETRIEVE) { List resultDesc; int dest; String intoName; char archiveMode; ObjectId intoRelationId; resultDesc = root_result_relation(parseRoot); if (! lispNullp(resultDesc)) { dest = CAtom(CAR(resultDesc)); if (dest == INTO) { /* ---------------- * create the "into" relation * * note: there is currently no way for the user to * specify the desired archive mode of the * "into" relation... * ---------------- */ intoName = CString(CADR(resultDesc)); archiveMode = 'n'; intoRelationId = heap_create(intoName, archiveMode, len, DEFAULT_SMGR, (struct attribute **)tupType); /* ---------------- * XXX rather than having to call setheapoverride(true) * and then back to false, we should change the * arguments to heap_open() instead.. * ---------------- */ setheapoverride(true); intoRelationDesc = heap_open(intoRelationId); local_heap_open(intoRelationDesc); setheapoverride(false); } else if (dest == INTOTEMP) { /* ---------------- * this is only for the parallel backends to save results in * temporary relations and collect them later * ---------------- */ intoRelationDesc = ExecCreatR(len, tupType, -1); if (! IsMaster) { SlaveTmpRelDescInit(); SlaveInfoP[MyPid].resultTmpRelDesc = CopyRelDescUsing(intoRelationDesc, SlaveTmpRelDescAlloc); } } } } set_es_into_relation_descriptor(estate, intoRelationDesc); /* ---------------- * return the type information.. * ---------------- */ attinfo = MakeList(lispInteger(len), tupType, -1); return attinfo; } /* ---------------------------------------------------------------- * EndPlan * * Cleans up the query plan -- closes files and free up storages * ---------------------------------------------------------------- */ List EndPlan(plan, estate) Plan plan; EState estate; { RelationInfo resultRelationInfo; List resultRelationInfoList; Relation intoRelationDesc; /* ---------------- * get information from state * ---------------- */ resultRelationInfo = get_es_result_relation_info(estate); resultRelationInfoList = get_es_result_relation_info_list(estate); intoRelationDesc = get_es_into_relation_descriptor(estate); /* ---------------- * shut down the query * ---------------- */ ExecEndNode(plan); /* ---------------- * destroy the executor "tuple" table. * ---------------- */ { TupleTable tupleTable = (TupleTable) get_es_tupleTable(estate); ExecDestroyTupleTable(tupleTable,true); /* was missing last arg */ set_es_tupleTable(estate, NULL); } /* ---------------- * close the result relations if necessary * ---------------- */ if (resultRelationInfo != NULL) { Relation resultRelationDesc; resultRelationDesc = get_ri_RelationDesc(resultRelationInfo); heap_close(resultRelationDesc); /* ---------------- * close indices on the result relation * ---------------- */ ExecCloseIndices(resultRelationInfo); } while (resultRelationInfoList != LispNil) { Relation resultRelationDesc; resultRelationInfo = (RelationInfo) CAR(resultRelationInfoList); resultRelationDesc = get_ri_RelationDesc(resultRelationInfo); heap_close(resultRelationDesc); resultRelationInfoList = CDR(resultRelationInfoList); } /* ---------------- * close the "into" relation if necessary * ---------------- */ if (intoRelationDesc != NULL) if (ParallelExecutorEnabled()) /* have to use shared buffer pool for paralle backends */ heap_close(intoRelationDesc); else local_heap_close(intoRelationDesc); /* ---------------- * return successfully * ---------------- */ return LispTrue; } /* ---------------------------------------------------------------- * ExecutePlan * * processes the query plan to retrieve 'tupleCount' tuples in the * direction specified. * Retrieves all tuples if tupleCount is 0 * *| XXX direction seems not to be used (I hope this isn't *| another example of lisp global scoping) -cim 8/7/89 * ---------------------------------------------------------------- */ TupleTableSlot ExecutePlan(estate, plan, parseTree, operation, numberTuples, direction, printfunc) EState estate; Plan plan; LispValue parseTree; int operation; int numberTuples; int direction; void (*printfunc)(); { Relation intoRelationDesc; JunkFilter junkfilter; TupleTableSlot slot; ItemPointer tupleid = NULL; ItemPointerData tuple_ctid; int current_tuple_count; TupleTableSlot result; /* result to return to ExecMain */ RuleLock locks; /* ---------------- * get information * ---------------- */ intoRelationDesc = get_es_into_relation_descriptor(estate); /* ---------------- * initialize local variables * ---------------- */ slot = NULL; current_tuple_count = 0; result = NULL; /* ---------------- * Set the direction. * ---------------- */ set_es_direction(estate, direction); /* ---------------- * Loop until we've processed the proper number * of tuples from the plan.. * ---------------- */ for(;;) { /* ---------------- * Execute the plan and obtain a tuple * ---------------- */ if (operation != NOTIFY) slot = ExecProcNode(plan); /* ---------------- * if the tuple is null, then we assume * there is nothing more to process so * we just return null... * ---------------- */ if (TupIsNull((Pointer) slot) && operation != NOTIFY) { result = NULL; break; } /* ---------------- * if we have a junk filter, then project a new * tuple with the junk removed. * * Store this new "clean" tuple in the place of the * original tuple. * * Also, extract all the junk ifnormation we need. * ---------------- */ if ((junkfilter = get_es_junkFilter(estate)) != (JunkFilter)NULL) { Datum datum; NameData attrName; HeapTuple newTuple; Boolean isNull; /* --------------- * extract the 'ctid' junk attribute. * --------------- */ if (operation == REPLACE || operation == DELETE) { strcpy(&(attrName), "ctid"); if (! ExecGetJunkAttribute(junkfilter, slot, &attrName, &datum, &isNull)) elog(WARN,"ExecutePlan: NO (junk) `ctid' was found!"); if (isNull) elog(WARN,"ExecutePlan: (junk) `ctid' is NULL!"); tupleid = (ItemPointer) DatumGetPointer(datum); tuple_ctid = *tupleid; /* make sure we don't free the ctid!! */ tupleid = &tuple_ctid; } /* --------------- * Find the "rule lock" info. (if it is there!) * --------------- */ strcpy(&(attrName), "lock"); if (! ExecGetJunkAttribute(junkfilter, slot, &attrName, &datum, &isNull)) locks = InvalidRuleLock; else { if (isNull) { /* * this is an error (I think...) * If we have a "lock" junk attribute, then it better * have a value! * NOTE: an empty lock i.e. "(numOfLocks: 0)"::lock * is a good non-null value, i.e. still 'isNull' must * be false */ elog(WARN, "ExecutePlan: (junk) rule lock is NULL!"); } locks = (RuleLock) DatumGetPointer(datum); /* * Make a copy of the lock, because when we call * amreplace, we will (probably?) free it and the * "locks" as returned by "ExecGetJunkAttribute" is * something that -for some reason- causes problems when * you try to pfree it! (believe me, I've tried it!). */ locks = prs2CopyLocks(locks); } /* --------------- * Finally create a new "clean" tuple with all junk attributes * removed and with the new "rule lock" info. * --------------- */ newTuple = ExecRemoveJunk(junkfilter, slot); slot = (TupleTableSlot) ExecStoreTuple((Pointer) newTuple, /* tuple to store */ (Pointer) slot, /* destination slot */ InvalidBuffer, /* this tuple has no buffer */ true); /* tuple should be pfreed */ } else { locks = InvalidRuleLock; } #ifdef PALLOC_DEBUG /* ---------------- * if debugging memory allocations, print memory dump * every so often * ---------------- */ { extern int query_tuple_count, query_tuple_max; if (query_tuple_max && query_tuple_max == ++query_tuple_count) { query_tuple_count = 0; dump_palloc_list("ExecutePlan", true); } } #endif PALLOC_DEBUG /* ---------------- * now that we have a tuple, do the appropriate thing * with it.. either return it to the user, add * it to a relation someplace, delete it from a * relation, or modify some of it's attributes. * ---------------- */ switch(operation) { case RETRIEVE: result = ExecRetrieve(slot, /* slot containing tuple */ printfunc, /* print function */ intoRelationDesc); /* "into" relation */ break; case APPEND: result = ExecAppend(slot, tupleid, estate, locks); break; case DELETE: result = ExecDelete(slot, tupleid, estate); break; case REPLACE: result = ExecReplace(slot, tupleid, estate, parseTree, locks); break; /* Total hack. I'm ignoring any accessor functions for Relation, RelationTupleForm, NameData. Assuming that NameData.data has offset 0. */ case NOTIFY: { RelationInfo rInfo = get_es_result_relation_info(estate); Relation rDesc = get_ri_RelationDesc(rInfo); Async_Notify(&rDesc->rd_rel->relname); result = NULL; current_tuple_count = 0; numberTuples = 1; elog(DEBUG, "ExecNotify %s",&rDesc->rd_rel->relname); } break; default: elog(DEBUG, "ExecutePlan: unknown operation in queryDesc"); result = NULL; break; } /* ---------------- * check our tuple count.. if we've returned the * proper number then return, else loop again and * process more tuples.. * ---------------- */ current_tuple_count += 1; if (numberTuples == current_tuple_count) break; } /* ---------------- * here, result is either a slot containing a tuple in the case * of a RETRIEVE or NULL otherwise. * ---------------- */ return result; } /* ---------------------------------------------------------------- * ExecRetrieve * * RETRIEVEs are easy.. we just pass the tuple to the appropriate * print function. The only complexity is when we do a * "retrieve into", in which case we insert the tuple into * the appropriate relation (note: this is a newly created relation * so we don't need to worry about indices or locks.) * ---------------------------------------------------------------- */ TupleTableSlot ExecRetrieve(slot, printfunc, intoRelationDesc) TupleTableSlot slot; void (*printfunc)(); Relation intoRelationDesc; { HeapTuple tuple; TupleDescriptor attrtype; /* ---------------- * get the heap tuple out of the tuple table slot * ---------------- */ tuple = (HeapTuple) ExecFetchTuple((Pointer) slot); attrtype = ExecSlotDescriptor((Pointer) slot); /* ---------------- * insert the tuple into the "into relation" * ---------------- */ if (intoRelationDesc != NULL) { if (ParallelExecutorEnabled()) /* for parallel executor, has to insert through shared buffer */ (void) heap_insert(intoRelationDesc, tuple, NULL); else (void) local_heap_insert(intoRelationDesc, tuple); IncrAppended(); } /* ---------------- * send the tuple to the front end (or the screen) * ---------------- */ (*printfunc)(tuple, attrtype); IncrRetrieved(); /* ---------------- * return the tuple slot back to ExecutePlan * ---------------- */ return slot; } /* ---------------------------------------------------------------- * ExecAppend * * APPENDs are trickier.. we have to insert the tuple into * the base relation, insert appropriate tuples into the * index relations and if we step on some locks along the way, * run the rule manager.. * ---------------------------------------------------------------- */ TupleTableSlot ExecAppend(slot, tupleid, estate, newlocks) TupleTableSlot slot; ItemPointer tupleid; EState estate; RuleLock newlocks; { HeapTuple tuple; RelationInfo resultRelationInfo; Relation resultRelationDesc; int numIndices; RuleLock locks; RuleLock indexLocks; HeapTuple returnedTuple; Buffer returnedBuffer; Prs2Status status; Prs2Stub ruleStubs; RuleLock stubLocks; HeapTuple tempTuple; ObjectId newId; /* ---------------- * get the heap tuple out of the tuple table slot * ---------------- */ tuple = (HeapTuple) ExecFetchTuple((Pointer) slot); /* ---------------- * get information on the result relation * ---------------- */ resultRelationInfo = get_es_result_relation_info(estate); resultRelationDesc = get_ri_RelationDesc(resultRelationInfo); /* ---------------- * process rules * ---------------- */ #ifdef EXEC_RULEMANAGER /* * NOTE: the check for rule stubs etc. is done by the rule manager. */ /* * Now call the rule manager, but only if necessary. */ if (prs2MustCallRuleManager(get_es_result_rel_ruleinfo(estate), (HeapTuple) NULL, InvalidBuffer, APPEND)) { status = prs2Main(estate, (RelationRuleInfo) NULL,/*only aplies to retrieves*/ APPEND, 0, resultRelationDesc, (HeapTuple) NULL, InvalidBuffer, tuple, InvalidBuffer, (HeapTuple) NULL, InvalidBuffer, (AttributeNumberPtr) NULL, 0, &returnedTuple, &returnedBuffer); if (status == PRS2_STATUS_INSTEAD) { /* * do not append the tuple */ return NULL; } else if (status == PRS2_STATUS_TUPLE_CHANGED) { /* ---------------- * store the returnedTuple in the slot and free the old * copy of the tuple. * ---------------- */ ExecStoreTuple((Pointer) returnedTuple, /* tuple to store */ (Pointer) slot, /* destination slot */ returnedBuffer, true); /* tuple should be pfreed */ tuple = returnedTuple; } } /*if prs2MustCallRuleManager*/ #endif EXEC_RULEMANAGER /* ---------------- * have to add code to preform unique checking here. * cim -12/1/89 * ---------------- */ /* ---------------- * If the user has explicitly specified some locks in the "replace" * statement, ignore all rule stubs & stuff and just put these * locks in the tuple. Obviously this is guaranteed to mess things * up if the user does not *really* know what he is doing, * i.e. if it is not me! _sp. * ---------------- */ if (newlocks != InvalidRuleLock) HeapTupleSetRuleLock(tuple, InvalidBuffer, newlocks); /* ---------------- * insert the tuple * ---------------- */ newId = heap_insert(resultRelationDesc, /* relation desc */ tuple, /* heap tuple */ 0); /* return: offset */ IncrAppended(); UpdateAppendOid(newId); /* ---------------- * process indices * * Note: heap_insert adds a new tuple to a relation. As a side * effect, the tupleid of the new tuple is placed in the new * tuple's t_ctid field. * ---------------- */ numIndices = get_ri_NumIndices(resultRelationInfo); if (numIndices > 0) { indexLocks = ExecInsertIndexTuples(tuple, &(tuple->t_ctid), estate); } /* ---------------- * XXX will we ever return anything meaningful? * ---------------- */ return NULL; } /* ---------------------------------------------------------------- * ExecDelete * * DELETE is like append, we delete the tuple and its * index tuples and then run the rule manager if necessary. * ---------------------------------------------------------------- */ TupleTableSlot ExecDelete(slot, tupleid, estate) TupleTableSlot slot; ItemPointer tupleid; EState estate; { HeapTuple tuple; RelationInfo resultRelationInfo; Relation resultRelationDesc; RuleLock locks; RuleLock indexLocks; Prs2Status status; HeapTuple oldTuple; Buffer oldTupleBuffer; HeapTuple rawTuple; Buffer rawTupleBuffer; HeapTuple dummyTuple; Buffer dummyBuffer; /* ---------------- * get the heap tuple out of the tuple table slot * ---------------- */ tuple = (HeapTuple) ExecFetchTuple((Pointer) slot); /* ---------------- * get the result relation information * ---------------- */ resultRelationInfo = get_es_result_relation_info(estate); resultRelationDesc = get_ri_RelationDesc(resultRelationInfo); #ifdef EXEC_RULEMANAGER /* ---------------- * process rules * ---------------- */ /* ---------------- * find the old & raw tuples * * NOTE: currently we do not store them anywhere! * therefore we will use the 'tupleid' to find the original * tuple (the "raw" tuple") and we will assume that "old" tuple * is the same as the "raw" tuple. * The only problem is the extra time spent to reactivate * backward chaining rules already activated but... what * else can we do ? */ rawTuple = heap_fetch(resultRelationDesc, NowTimeQual, tupleid, &rawTupleBuffer); oldTuple = rawTuple; oldTupleBuffer = rawTupleBuffer; /* ---------------- * XXX!! * NOTE: in the current implementation where only RelationLevel * locks are used, the rule manager must look in pg_relation * to find the locks of "oldTuple". */ if (prs2MustCallRuleManager(get_es_result_rel_ruleinfo(estate), oldTuple, oldTupleBuffer, DELETE)) { status = prs2Main(estate, (RelationRuleInfo) NULL,/*only aplies to retrieves*/ DELETE, 0, /* userid */ resultRelationDesc, oldTuple, oldTupleBuffer, (HeapTuple) NULL, InvalidBuffer, rawTuple, rawTupleBuffer, (AttributeNumberPtr) NULL, 0, &dummyTuple, &dummyBuffer); if (status == PRS2_STATUS_INSTEAD) { /* * there was an instead rule, do not do the delete... */ return NULL; } } /*if prs2MustCallRuleManager*/ if (BufferIsValid(rawTupleBuffer)) ReleaseBuffer(rawTupleBuffer); #endif EXEC_RULEMANAGER /* ---------------- * delete the tuple * ---------------- */ locks = heap_delete(resultRelationDesc, /* relation desc */ tupleid); /* item pointer to tuple */ IncrDeleted(); /* ---------------- * Note: Normally one would think that we have to * delete index tuples associated with the * heap tuple now.. * * ... but in POSTGRES, we have no need to do this * because the vaccuum daemon automatically * opens an index scan and deletes index tuples * when it finds deleted heap tuples. -cim 9/27/89 * ---------------- */ /* ---------------- * XXX will we ever return anything meaningful? * ---------------- */ return NULL; } /* ---------------------------------------------------------------- * ExecReplace * * note: we can't run replace queries with transactions * off because replaces are actually appends and our * scan will mistakenly loop forever, replacing the tuple * it just appended.. This should be fixed but until it * is, we don't want to get stuck in an infinite loop * which corrupts your database.. * ---------------------------------------------------------------- */ TupleTableSlot ExecReplace(slot, tupleid, estate, parseTree, newlocks) TupleTableSlot slot; ItemPointer tupleid; EState estate; LispValue parseTree; RuleLock newlocks; { HeapTuple tuple; RelationInfo resultRelationInfo; Relation resultRelationDesc; RuleLock locks; RuleLock indexLocks; int numIndices; HeapTuple rawTuple; Buffer rawTupleBuffer = InvalidBuffer; HeapTuple oldTuple; Buffer oldTupleBuffer; HeapTuple changedTuple; Buffer changedTupleBuffer; AttributeNumberPtr attributeArray; AttributeNumber numberOfReplacedAttributes; LispValue targetList; LispValue t; Resdom resdom; Prs2Status status; Prs2Stub ruleStubs; RuleLock stubLocks; HeapTuple tempTuple; /* ---------------- * abort the operation if not running transactions * ---------------- */ if (IsBootstrapProcessingMode()) { elog(DEBUG, "ExecReplace: replace can't run without transactions"); return NULL; } /* ---------------- * get the heap tuple out of the tuple table slot * ---------------- */ tuple = (HeapTuple) ExecFetchTuple((Pointer) slot); /* ---------------- * get the result relation information * ---------------- */ resultRelationInfo = get_es_result_relation_info(estate); resultRelationDesc = get_ri_RelationDesc(resultRelationInfo); /* ---------------- * process rules * ---------------- */ #ifdef EXEC_RULEMANAGER /* ---------------- * find the old & raw tuples * * NOTE: currently we do not store them anywhere! * therefore we will use the 'tupleid' to find the original * tuple (the "raw" tuple") and we will assume that "old" tuple * is the same as the "raw" tuple. * The only problem is the extra time spent to reactivate * backward chaining rules already activated but... what * else can we do ? */ rawTuple = heap_fetch(resultRelationDesc, NowTimeQual, tupleid, &rawTupleBuffer); oldTuple = rawTuple; oldTupleBuffer = rawTupleBuffer; if (prs2MustCallRuleManager(get_es_result_rel_ruleinfo(estate), oldTuple, oldTupleBuffer, REPLACE)) { /* ---------------- * find the attributes replaced... */ attributeArray = (AttributeNumberPtr) palloc( sizeof(AttributeNumber) * RelationGetNumberOfAttributes(resultRelationDesc)); numberOfReplacedAttributes = 0; targetList = parse_targetlist(parseTree); foreach (t, targetList) { resdom = (Resdom) CAR(CAR(t)); attributeArray[numberOfReplacedAttributes] = get_resno(resdom); numberOfReplacedAttributes += 1; } status = prs2Main(estate, (RelationRuleInfo) NULL,/*only aplies to retrieves*/ REPLACE, 0, /* user id */ resultRelationDesc, oldTuple, oldTupleBuffer, tuple, InvalidBuffer, rawTuple, oldTupleBuffer, attributeArray, numberOfReplacedAttributes, &changedTuple, &changedTupleBuffer); if (status == PRS2_STATUS_INSTEAD) { /* * An instead rule has been found, do not do the replace... */ return NULL; } if (status == PRS2_STATUS_TUPLE_CHANGED) { ExecStoreTuple((Pointer) changedTuple, /* tuple to store */ (Pointer) slot, /* destination slot */ InvalidBuffer, /* this tuple has no buffer */ true); /* tuple should be pfreed */ tuple = changedTuple; } } /*if prs2MustCallRuleManager */ if (BufferIsValid(oldTupleBuffer)) ReleaseBuffer(oldTupleBuffer); #endif EXEC_RULEMANAGER /* ---------------- * have to add code to preform unique checking here. * in the event of unique tuples, this becomes a deletion * of the original tuple affected by the replace. * cim -12/1/89 * ---------------- */ /* ---------------- * If the user has explicitly specified some locks in the "replace" * statement, ignore all rule stubs & stuff and just put these * locks in the tuple. Obviously this is guaranteed to mess things * up if the user does not *really* know what he is doing, * i.e. if it is not me! _sp. * ---------------- */ if (newlocks != InvalidRuleLock) HeapTupleSetRuleLock(tuple, InvalidBuffer, newlocks); /* ---------------- * replace the heap tuple * ---------------- */ locks = heap_replace(resultRelationDesc, /* relation desc */ tupleid, /* item ptr of tuple to replace */ tuple); /* replacement heap tuple */ IncrReplaced(); /* ---------------- * Note: instead of having to update the old index tuples * associated with the heap tuple, all we do is form * and insert new index tuples.. This is because * replaces are actually deletes and inserts and * index tuple deletion is done automagically by * the vaccuum deamon.. All we do is insert new * index tuples. -cim 9/27/89 * ---------------- */ /* ---------------- * process indices * * heap_replace updates a tuple in the base relation by invalidating * it and then appending a new tuple to the relation. As a side * effect, the tupleid of the new tuple is placed in the new * tuple's t_ctid field. So we now insert index tuples using * the new tupleid stored there. * ---------------- */ numIndices = get_ri_NumIndices(resultRelationInfo); if (numIndices > 0) { indexLocks = ExecInsertIndexTuples(tuple, &(tuple->t_ctid), estate); } /* ---------------- * XXX will we ever return anything meaningful? * ---------------- */ return NULL; }