本节简单解释了PostgreSQL NOT IN在执行时为何会出现时快时慢的现象。 测试数据如下： ``` [local]:5432 pg12@testdb=# select count(*) from tbl; count ------- 1 (1 row) Time: 6.009 ms [local]:5432 pg12@testdb=# select count(*) from t_big_null; count ---------- 10000001 (1 row) Time: 633.248 ms [local]:5432 pg12@testdb=# \d tbl Table "public.tbl" Column | Type | Collation | Nullable | Default --------+---------+-----------+----------+--------- id | integer | | not null | value | integer | | not null | Indexes: "tbl_pkey" PRIMARY KEY, btree (id) Rules: rule_tbl_update AS ON INSERT TO tbl WHERE (EXISTS ( SELECT tbl_1.id, tbl_1.value FROM tbl tbl_1 WHERE tbl_1.id = new.id)) DO INSTEAD UPDATE tbl SET value = tbl.value + 1 WHERE tbl.id = new.id [local]:5432 pg12@testdb=# \d t_big_null Table "public.t_big_null" Column | Type | Collation | Nullable | Default --------+---------+-----------+----------+--------- id | integer | | | [local]:5432 pg12@testdb=# ``` **注意tbl表只有一行数据(id = 1)，而t_big_null表在插入"id = 1"这一行时有意放在最后才插入** ``` truncate table t_big_null; insert into t_big_null select generate_series(2,10000000); insert into t_big_null values(1); ``` ### 一、数据结构 **SubPlanState** 子计划运行期状态 ``` /* ---------------- * SubPlanState node * ---------------- */ typedef struct SubPlanState { NodeTag type; SubPlan *subplan; /* expression plan node */ struct PlanState *planstate; /* subselect plan's state tree */ struct PlanState *parent; /* parent plan node's state tree */ ExprState *testexpr; /* 组合表达式状态；state of combining expression */ List *args; /* 参数表达式状态；states of argument expression(s) */ HeapTuple curTuple; /* subplan最近的元组；copy of most recent tuple from subplan */ Datum curArray; /* most recent array from ARRAY() subplan */ /* these are used when hashing the subselect's output: */ TupleDesc descRight; /* 投影后的子查询描述符；subselect desc after projection */ ProjectionInfo *projLeft; /* for projecting lefthand exprs */ ProjectionInfo *projRight; /* for projecting subselect output */ TupleHashTable hashtable; /* hash table for no-nulls subselect rows */ TupleHashTable hashnulls; /* hash table for rows with null(s) */ bool havehashrows; /* true if hashtable is not empty */ bool havenullrows; /* true if hashnulls is not empty */ MemoryContext hashtablecxt; /* memory context containing hash tables */ MemoryContext hashtempcxt; /* temp memory context for hash tables */ ExprContext *innerecontext; /* econtext for computing inner tuples */ AttrNumber *keyColIdx; /* control data for hash tables */ Oid *tab_eq_funcoids; /* equality func oids for table * datatype(s) */ Oid *tab_collations; /* collations for hash and comparison */ FmgrInfo *tab_hash_funcs; /* hash functions for table datatype(s) */ FmgrInfo *tab_eq_funcs; /* equality functions for table datatype(s) */ FmgrInfo *lhs_hash_funcs; /* hash functions for lefthand datatype(s) */ FmgrInfo *cur_eq_funcs; /* equality functions for LHS vs. table */ ExprState *cur_eq_comp; /* equality comparator for LHS vs. table */ } SubPlanState; ``` **SubPlan** 子查询计划 ``` /* * SubPlan - executable expression node for a subplan (sub-SELECT) * * The planner replaces SubLink nodes in expression trees with SubPlan * nodes after it has finished planning the subquery. SubPlan references * a sub-plantree stored in the subplans list of the toplevel PlannedStmt. * (We avoid a direct link to make it easier to copy expression trees * without causing multiple processing of the subplan.) * 查询规划器在完成子查询的规划后使用SubPlan节点替换表达式树中的SubLink节点。 * SubPlan引用了存储在高层PlannedStmt中的subplans链表中的sub-plantree。 * （避免使用直接链接，从而使得拷贝表达式树相对比较简单） * * In an ordinary subplan, testexpr points to an executable expression * (OpExpr, an AND/OR tree of OpExprs, or RowCompareExpr) for the combining * operator(s); the left-hand arguments are the original lefthand expressions, * and the right-hand arguments are PARAM_EXEC Param nodes representing the * outputs of the sub-select. (NOTE: runtime coercion functions may be * inserted as well.) This is just the same expression tree as testexpr in * the original SubLink node, but the PARAM_SUBLINK nodes are replaced by * suitably numbered PARAM_EXEC nodes. * 常规情况下，testexpr指向用于组合操作的可执行表达式（OpExpr、OpExprs的AND/OR树或者RowCompareExpr）； * 左参数是原始的左表达式，右参数是PARAM_EXEC参数节点用以表示子查询的输出。 * 与原始SubLink节点的testexpr具有相同的表达式树，但PARAM_SUBLINK节点则使用合适的已编号PARAM_EXEC节点替代。 * * If the sub-select becomes an initplan rather than a subplan, the executable * expression is part of the outer plan's expression tree (and the SubPlan * node itself is not, but rather is found in the outer plan's initPlan * list). In this case testexpr is NULL to avoid duplication. * 如果子查询成了initplan而不是subplan，可执行的表达式是外层plan表达式树的一部分。 * 这种情况下，testexpr为NULL以避免重复。 * * The planner also derives lists of the values that need to be passed into * and out of the subplan. Input values are represented as a list "args" of * expressions to be evaluated in the outer-query context (currently these * args are always just Vars, but in principle they could be any expression). * The values are assigned to the global PARAM_EXEC params indexed by parParam * (the parParam and args lists must have the same ordering). setParam is a * list of the PARAM_EXEC params that are computed by the sub-select, if it * is an initplan; they are listed in order by sub-select output column * position. (parParam and setParam are integer Lists, not Bitmapsets, * because their ordering is significant.) * 规划器还派生了需要传入和传出子计划的值的链表。 * 输入值标识位表达式的“args”链表，在外层查询上下文中进行解析。 * （这些args通常是Vars，但原则上它们可以是任意表达式） * 这些值以parParam为索引给全局PARAM_EXEC参数赋值。 * setParam是PARAM_EXEC参数链表，通过子查询（如为initplan）计算所得。 * 它们按子查询输出列的位置进行排序组织为链表形式。 * （parParam和setParam是整型链表，而不是Bitmapsets链表） * * Also, the planner computes startup and per-call costs for use of the * SubPlan. Note that these include the cost of the subquery proper, * evaluation of the testexpr if any, and any hashtable management overhead. * 同时，规划器计算SubPlan启动和每次调用的成本。注意：包括子查询正常解析testexpr的成本以及哈希表管理成本。 */ typedef struct SubPlan { Expr xpr;//表达式 /* Fields copied from original SubLink: */ //从SubLink中拷贝而来 SubLinkType subLinkType; /* see above */ /* The combining operators, transformed to an executable expression: */ //组合操作符,转换为可执行的表达式 Node *testexpr; /* OpExpr or RowCompareExpr expression tree */ List *paramIds; /* 参数IDs;IDs of Params embedded in the above */ /* Identification of the Plan tree to use: */ //Plan tree标识 int plan_id; /* Index (from 1) in PlannedStmt.subplans */ /* Identification of the SubPlan for EXPLAIN and debugging purposes: */ //EXPLAIN和debug目的的SubPlan标识 char *plan_name; /* A name assigned during planning */ /* Extra data useful for determining subplan's output type: */ //用于确定subplan输出类型的额外信息 Oid firstColType; /* subplan结果的第一个列类型;Type of first column of subplan result */ int32 firstColTypmod; /* 第一列的Typmod;Typmod of first column of subplan result */ Oid firstColCollation; /* 第一列的Collation;Collation of first column of subplan * result */ /* Information about execution strategy: */ //执行阶段的相关信息 bool useHashTable; /* 是否使用哈希表存储子查询输出;true to store subselect output in a hash * table (implies we are doing "IN") */ bool unknownEqFalse; /* 如OK为T,如为未知则为F;快速处理null值;true if it's okay to return FALSE when the * spec result is UNKNOWN; this allows much * simpler handling of null values */ bool parallel_safe; /* 是否并行安全?is the subplan parallel-safe? */ /* Note: parallel_safe does not consider contents of testexpr or args */ /* Information for passing params into and out of the subselect: */ //用于给子查询传入和传出参数的信息 /* setParam and parParam are lists of integers (param IDs) */ //setParam和parParam是整型链表(param IDs) List *setParam; /* initplan subqueries have to set these * Params for parent plan */ List *parParam; /* indices of input Params from parent plan */ List *args; /* 以parParam值进行传递的表达式；exprs to pass as parParam values */ /* Estimated execution costs: */ //估算执行成本 Cost startup_cost; /* one-time setup cost */ Cost per_call_cost; /* cost for each subplan evaluation */ } SubPlan; ``` **SubLinkType** SubLink类型 ``` /* * SubLink * * A SubLink represents a subselect appearing in an expression, and in some * cases also the combining operator(s) just above it. The subLinkType * indicates the form of the expression represented: * EXISTS_SUBLINK EXISTS(SELECT ...) * ALL_SUBLINK (lefthand) op ALL (SELECT ...) * ANY_SUBLINK (lefthand) op ANY (SELECT ...) * ROWCOMPARE_SUBLINK (lefthand) op (SELECT ...) * EXPR_SUBLINK (SELECT with single targetlist item ...) * MULTIEXPR_SUBLINK (SELECT with multiple targetlist items ...) * ARRAY_SUBLINK ARRAY(SELECT with single targetlist item ...) * CTE_SUBLINK WITH query (never actually part of an expression) * 我们使用SubLink表示在表达式中出现的子查询，在某些情况下组合操作符会出现在SubLink之上。 * subLinkType表示表达式的形式： * EXISTS_SUBLINK EXISTS(SELECT ...) * ALL_SUBLINK (lefthand) op ALL (SELECT ...) * ANY_SUBLINK (lefthand) op ANY (SELECT ...) * ROWCOMPARE_SUBLINK (lefthand) op (SELECT ...) * EXPR_SUBLINK (SELECT with single targetlist item ...) * MULTIEXPR_SUBLINK (SELECT with multiple targetlist items ...) * ARRAY_SUBLINK ARRAY(SELECT with single targetlist item ...) * CTE_SUBLINK WITH query (never actually part of an expression) * * For ALL, ANY, and ROWCOMPARE, the lefthand is a list of expressions of the * same length as the subselect's targetlist. ROWCOMPARE will *always* have * a list with more than one entry; if the subselect has just one target * then the parser will create an EXPR_SUBLINK instead (and any operator * above the subselect will be represented separately). * ROWCOMPARE, EXPR, and MULTIEXPR require the subselect to deliver at most * one row (if it returns no rows, the result is NULL). * ALL, ANY, and ROWCOMPARE require the combining operators to deliver boolean * results. ALL and ANY combine the per-row results using AND and OR * semantics respectively. * ARRAY requires just one target column, and creates an array of the target * column's type using any number of rows resulting from the subselect. * 对于ALL,ANY和ROWCOMPARE,左操作符是与子查询目标链表长度一致的表达式链表。 * ROWCOMPARE通常有超过一个条目的链表；如果子查询刚好只有一个目标列，那么解析器会创建EXPR_SUBLINK * （同时所有在子查询之上的操作符会单独表示） * ROWCOMPARE, EXPR, 和MULTIEXPR要求子查询至少输出一行（如返回0行，则结果为NULL）。 * ALL,ANY和ROWCOMPARE要求组合操作符输出布尔型结果。 * ALL/ANY使用AND/OR语义来组合每一行的结果。 * * SubLink is classed as an Expr node, but it is not actually executable; * it must be replaced in the expression tree by a SubPlan node during * planning. * SubLink归类为Expr节点，但实际上并不是可执行的，必须在计划阶段通过SubPlan替代。 * * NOTE: in the raw output of gram.y, testexpr contains just the raw form * of the lefthand expression (if any), and operName is the String name of * the combining operator. Also, subselect is a raw parsetree. During parse * analysis, the parser transforms testexpr into a complete boolean expression * that compares the lefthand value(s) to PARAM_SUBLINK nodes representing the * output columns of the subselect. And subselect is transformed to a Query. * This is the representation seen in saved rules and in the rewriter. * 注意：在gram.y的裸输出中，testexpr只包含左表达式的裸形式，operName是组合操作符的字符串名称。 * 同时，子查询是裸parsetree。在解析分析期间， * 解析器转换testexpr为完整的布尔表达式用于比较左操作符值与PARAM_SUBLINK节点所代表的子查询输出列值。 * 子查询会转换为Query结构体。 * 在已存储的规则和重写时可见的表示形式。 * * In EXISTS, EXPR, MULTIEXPR, and ARRAY SubLinks, testexpr and operName * are unused and are always null. * 在EXISTS/EXPR/MULTEXPR/ARRAY SubLinks中，testexpr和operName不再使用通常是NULL值。 * * subLinkId is currently used only for MULTIEXPR SubLinks, and is zero in * other SubLinks. This number identifies different multiple-assignment * subqueries within an UPDATE statement's SET list. It is unique only * within a particular targetlist. The output column(s) of the MULTIEXPR * are referenced by PARAM_MULTIEXPR Params appearing elsewhere in the tlist. * subLinkId当前只用于MULTIEXPR，在其他SubLinks中取值为0. * 该数字标识了在UPDATE语句SET链表中不同的多个赋值子查询。 * 只有在特定的targetlist内是唯一的。 * 出现在tlist其他地方的PARAM_MULTIEXPR参数依赖于MULTIEXPR的输出列。 * * The CTE_SUBLINK case never occurs in actual SubLink nodes, but it is used * in SubPlans generated for WITH subqueries. * CTE_SUBLINK不会出现在实际的SubLink节点中，但用于WITH子查询所产生的SubPlans中。 */ typedef enum SubLinkType { EXISTS_SUBLINK, ALL_SUBLINK, ANY_SUBLINK, ROWCOMPARE_SUBLINK, EXPR_SUBLINK, MULTIEXPR_SUBLINK, ARRAY_SUBLINK, CTE_SUBLINK /* 仅用于SubPlans中；for SubPlans only */ } SubLinkType; ``` **SubLink** SubLink结构体 ``` typedef struct SubLink { Expr xpr; SubLinkType subLinkType; /* see above */ int subLinkId; /* ID (1..n); 0 if not MULTIEXPR */ Node *testexpr; /* outer-query test for ALL/ANY/ROWCOMPARE */ List *operName; /* originally specified operator name */ Node *subselect; /* subselect as Query* or raw parsetree */ int location; /* token location, or -1 if unknown */ } SubLink; ``` **MaterialState** Material状态 ``` /* ---------------- * MaterialState information * * materialize nodes are used to materialize the results * of a subplan into a temporary file. * materialize节点用于物化subplan的结果为临时文件。 * * ss.ss_ScanTupleSlot refers to output of underlying plan. * ss.ss_ScanTupleSlot指向underlyling plan的输出（subplan） * ---------------- */ typedef struct MaterialState { ScanState ss; /* its first field is NodeTag */ int eflags; /* 传递给tuplestore的capability标记；capability flags to pass to tuplestore */ bool eof_underlying; /* 已经到达underlying plan的末尾？reached end of underlying plan? */ Tuplestorestate *tuplestorestate; } MaterialState; ``` **Tuplestorestate** Tuplestore相关操作的私有状态。 ``` /* * Possible states of a Tuplestore object. These denote the states that * persist between calls of Tuplestore routines. */ typedef enum { TSS_INMEM, /* Tuples still fit in memory */ TSS_WRITEFILE, /* Writing to temp file */ TSS_READFILE /* Reading from temp file */ } TupStoreStatus; /* * Private state of a Tuplestore operation. */ struct Tuplestorestate { TupStoreStatus status; /* 状态枚举值；enumerated value as shown above */ int eflags; /* capability flags (OR of pointers' flags) */ bool backward; /* store extra length words in file? */ bool interXact; /* keep open through transactions? */ bool truncated; /* tuplestore_trim has removed tuples? */ int64 availMem; /* remaining memory available, in bytes */ int64 allowedMem; /* total memory allowed, in bytes */ int64 tuples; /* number of tuples added */ BufFile *myfile; /* underlying file, or NULL if none */ MemoryContext context; /* memory context for holding tuples */ ResourceOwner resowner; /* resowner for holding temp files */ /* * These function pointers decouple the routines that must know what kind * of tuple we are handling from the routines that don't need to know it. * They are set up by the tuplestore_begin_xxx routines. * * (Although tuplestore.c currently only supports heap tuples, I've copied * this part of tuplesort.c so that extension to other kinds of objects * will be easy if it's ever needed.) * * Function to copy a supplied input tuple into palloc'd space. (NB: we * assume that a single pfree() is enough to release the tuple later, so * the representation must be "flat" in one palloc chunk.) state->availMem * must be decreased by the amount of space used. */ void *(*copytup) (Tuplestorestate *state, void *tup); /* * Function to write a stored tuple onto tape. The representation of the * tuple on tape need not be the same as it is in memory; requirements on * the tape representation are given below. After writing the tuple, * pfree() it, and increase state->availMem by the amount of memory space * thereby released. */ void (*writetup) (Tuplestorestate *state, void *tup); /* * Function to read a stored tuple from tape back into memory. 'len' is * the already-read length of the stored tuple. Create and return a * palloc'd copy, and decrease state->availMem by the amount of memory * space consumed. */ void *(*readtup) (Tuplestorestate *state, unsigned int len); /* * This array holds pointers to tuples in memory if we are in state INMEM. * In states WRITEFILE and READFILE it's not used. * * When memtupdeleted > 0, the first memtupdeleted pointers are already * released due to a tuplestore_trim() operation, but we haven't expended * the effort to slide the remaining pointers down. These unused pointers * are set to NULL to catch any invalid accesses. Note that memtupcount * includes the deleted pointers. */ void **memtuples; /* array of pointers to palloc'd tuples */ int memtupdeleted; /* the first N slots are currently unused */ int memtupcount; /* number of tuples currently present */ int memtupsize; /* allocated length of memtuples array */ bool growmemtuples; /* memtuples' growth still underway? */ /* * These variables are used to keep track of the current positions. * * In state WRITEFILE, the current file seek position is the write point; * in state READFILE, the write position is remembered in writepos_xxx. * (The write position is the same as EOF, but since BufFileSeek doesn't * currently implement SEEK_END, we have to remember it explicitly.) */ TSReadPointer *readptrs; /* array of read pointers */ int activeptr; /* index of the active read pointer */ int readptrcount; /* number of pointers currently valid */ int readptrsize; /* allocated length of readptrs array */ int writepos_file; /* file# (valid if READFILE state) */ off_t writepos_offset; /* offset (valid if READFILE state) */ }; #define COPYTUP(state,tup) ((*(state)->copytup) (state, tup)) #define WRITETUP(state,tup) ((*(state)->writetup) (state, tup)) #define READTUP(state,len) ((*(state)->readtup) (state, len)) #define LACKMEM(state) ((state)->availMem < 0) #define USEMEM(state,amt) ((state)->availMem -= (amt)) #define FREEMEM(state,amt) ((state)->availMem += (amt)) ``` **TSReadPointer** tuplestore读指针 ``` /* * Possible states of a Tuplestore object. These denote the states that * persist between calls of Tuplestore routines. */ typedef enum { TSS_INMEM, /* Tuples still fit in memory */ TSS_WRITEFILE, /* Writing to temp file */ TSS_READFILE /* Reading from temp file */ } TupStoreStatus; /* * State for a single read pointer. If we are in state INMEM then all the * read pointers' "current" fields denote the read positions. In state * WRITEFILE, the file/offset fields denote the read positions. In state * READFILE, inactive read pointers have valid file/offset, but the active * read pointer implicitly has position equal to the temp file's seek position. * * Special case: if eof_reached is true, then the pointer's read position is * implicitly equal to the write position, and current/file/offset aren't * maintained. This way we need not update all the read pointers each time * we write. */ typedef struct { int eflags; /* capability flags */ bool eof_reached; /* read has reached EOF */ int current; /* next array index to read */ int file; /* temp file# */ off_t offset; /* byte offset in file */ } TSReadPointer; ``` ### 二、源码解读 NOT IN在实际执行时会转换为ALL_SUBLINK，执行的快慢取决于什么时候会碰到符合条件的记录，一旦遇到马上返回。因此，SQL的执行时间与数据表的扫描顺序有紧密的关系，符合条件的行越早出现，程序越早返回，需要的时间越短。 相关代码如下： ``` ... //解析表达式 rowresult = ExecEvalExprSwitchContext(node->testexpr, econtext, &rownull); if (subLinkType == ANY_SUBLINK) { //ANY : 使用OR语义组合 /* combine across rows per OR semantics */ if (rownull) *isNull = true; else if (DatumGetBool(rowresult)) { result = BoolGetDatum(true); *isNull = false; break; /* needn't look at any more rows */ } } ``` 如上代码所示，在全表扫描物化的时候一旦textexpr表达式测试到其中一行满足条件，那么就会马上退出循环，而这一行如果非常幸运的出现在扫描的最开始的地方，那执行时间将会很快（扫描几个数据块 vs 全表扫描）。 新建一张表，插入2条记录，其中id = 2的行出现在block编号最小的地方，而id = 1出现在block编号最大的地方，这时候pg就会出现时快时慢的情况，两者相差3个数量级。 ``` [local]:5432 pg12@testdb=# create table tbl3(id int); CREATE TABLE Time: 1.852 ms [local]:5432 pg12@testdb=# insert into tbl3 values(1); INSERT 0 1 Time: 1.276 ms [local]:5432 pg12@testdb=# insert into tbl3 values(2); INSERT 0 1 Time: 1.089 ms [local]:5432 pg12@testdb=# select * from tbl3 where id not in (select b.id from t_big_null b); id ---- (0 rows) Time: 3.676 ms [local]:5432 pg12@testdb=# select * from tbl3 where id not in (select b.id from t_big_null b); id ---- (0 rows) Time: 4925.893 ms (00:04.926) [local]:5432 pg12@testdb=# select * from tbl3 where id not in (select b.id from t_big_null b); id ---- (0 rows) Time: 2.858 ms [local]:5432 pg12@testdb=# select * from tbl3 where id not in (select b.id from t_big_null b); id ---- (0 rows) Time: 4588.436 ms (00:04.588) [local]:5432 pg12@testdb=# select * from tbl3 where id not in (select b.id from t_big_null b); id ---- (0 rows) Time: 1.896 ms [local]:5432 pg12@testdb=# select * from tbl3 where id not in (select b.id from t_big_null b); id ---- (0 rows) Time: 4653.525 ms (00:04.654) [local]:5432 pg12@testdb=# ``` **ExecScanSubPlan** ``` /* * ExecScanSubPlan: default case where we have to rescan subplan each time * 默认情况下每次都不得不重新扫描subplan */ static Datum ExecScanSubPlan(SubPlanState *node, ExprContext *econtext, bool *isNull) { SubPlan *subplan = node->subplan;//子计划 PlanState *planstate = node->planstate;//计划运行期状态 SubLinkType subLinkType = subplan->subLinkType;//子链接类型 MemoryContext oldcontext;//原内存上下文 TupleTableSlot *slot;//元组slot Datum result;//结果指针 bool found = false; /* 如找到至少一个元组,则返回T;true if got at least one subplan tuple */ ListCell *pvar;//临时变量 ListCell *l;//临时变量 ArrayBuildStateAny *astate = NULL;// /* * MULTIEXPR subplans, when "executed", just return NULL; but first we * mark the subplan's output parameters as needing recalculation. (This * is a bit of a hack: it relies on the subplan appearing later in its * targetlist than any of the referencing Params, so that all the Params * have been evaluated before we re-mark them for the next evaluation * cycle. But in general resjunk tlist items appear after non-resjunk * ones, so this should be safe.) Unlike ExecReScanSetParamPlan, we do * *not* set bits in the parent plan node's chgParam, because we don't * want to cause a rescan of the parent. * * MULTIEXPR处理逻辑 */ if (subLinkType == MULTIEXPR_SUBLINK) { EState *estate = node->parent->state; foreach(l, subplan->setParam) { int paramid = lfirst_int(l); ParamExecData *prm = &(estate->es_param_exec_vals[paramid]); prm->execPlan = node; } *isNull = true; return (Datum) 0; } /* Initialize ArrayBuildStateAny in caller's context, if needed */ //数组 if (subLinkType == ARRAY_SUBLINK) astate = initArrayResultAny(subplan->firstColType, CurrentMemoryContext, true); /* * We are probably in a short-lived expression-evaluation context. Switch * to the per-query context for manipulating the child plan's chgParam, * calling ExecProcNode on it, etc. */ //切换上下文 oldcontext = MemoryContextSwitchTo(econtext->ecxt_per_query_memory); /* * Set Params of this plan from parent plan correlation values. (Any * calculation we have to do is done in the parent econtext, since the * Param values don't need to have per-query lifetime.) */ //通过父计划相关值中设置子计划参数 Assert(list_length(subplan->parParam) == list_length(node->args)); forboth(l, subplan->parParam, pvar, node->args) { int paramid = lfirst_int(l); ParamExecData *prm = &(econtext->ecxt_param_exec_vals[paramid]); prm->value = ExecEvalExprSwitchContext((ExprState *) lfirst(pvar), econtext, &(prm->isnull)); planstate->chgParam = bms_add_member(planstate->chgParam, paramid); } /* * Now that we've set up its parameters, we can reset the subplan. */ //执行ReScan //Reset a plan node so that its output can be re-scanned. ExecReScan(planstate); /* * For all sublink types except EXPR_SUBLINK and ARRAY_SUBLINK, the result * is boolean as are the results of the combining operators. We combine * results across tuples (if the subplan produces more than one) using OR * semantics for ANY_SUBLINK or AND semantics for ALL_SUBLINK. * (ROWCOMPARE_SUBLINK doesn't allow multiple tuples from the subplan.) * NULL results from the combining operators are handled according to the * usual SQL semantics for OR and AND. The result for no input tuples is * FALSE for ANY_SUBLINK, TRUE for ALL_SUBLINK, NULL for * ROWCOMPARE_SUBLINK. * 除EXPR_SUBLINK和ARRAY_SUBLINK外的所有sublink,结果是布尔值(组合运算符的结果). * PG通过跨元组(如子计划产生多个元组)合并结果,对于ANY_SUBLINK使用OR语义,ALL_SUBLINK则使用AND语义. * (ROWCOMPARE_SUBLINK不允许子计划返回多个元组) * 从组合操作符中返回的NULL遵循SQL中的OR和AND语义. * 如没有输入元组,ANY_SUBLINK为FALSE,ALL_SUBLINK为TRUE,ROWCOMPARE_SUBLINK为NULL. * * For EXPR_SUBLINK we require the subplan to produce no more than one * tuple, else an error is raised. If zero tuples are produced, we return * NULL. Assuming we get a tuple, we just use its first column (there can * be only one non-junk column in this case). * 对于EXPR_SUBLINK,需要subplan产生不超过一个元组,否则报错.如果没有元组产生,返回NULL. * 假定获取到一个元组,则使用第一个列(这种情况下只有一个non-junk列). * * For ARRAY_SUBLINK we allow the subplan to produce any number of tuples, * and form an array of the first column's values. Note in particular * that we produce a zero-element array if no tuples are produced (this is * a change from pre-8.3 behavior of returning NULL). * 对于ARRAY_SUBLINK,允许subplan产生任意数目的元组,使用第一个列值组成数组. * 特别注意的是如没有元组产生则产生0个元素的数组(8.3以前是返回NULL). */ result = BoolGetDatum(subLinkType == ALL_SUBLINK);//ALL为T,否则为F *isNull = false; for (slot = ExecProcNode(planstate); !TupIsNull(slot); slot = ExecProcNode(planstate))//循环获取元组,直至没有元组为NULL(即已完成) { //元组描述符 TupleDesc tdesc = slot->tts_tupleDescriptor; Datum rowresult;//结果 bool rownull;//是否为空? int col;//列计数器 ListCell *plst;//临时变量 if (subLinkType == EXISTS_SUBLINK)//EXISTS { found = true; result = BoolGetDatum(true); break; } if (subLinkType == EXPR_SUBLINK)//EXPR表达式 { /* cannot allow multiple input tuples for EXPR sublink */ if (found) ereport(ERROR, (errcode(ERRCODE_CARDINALITY_VIOLATION), errmsg("more than one row returned by a subquery used as an expression"))); found = true; /* * We need to copy the subplan's tuple in case the result is of * pass-by-ref type --- our return value will point into this * copied tuple! Can't use the subplan's instance of the tuple * since it won't still be valid after next ExecProcNode() call. * node->curTuple keeps track of the copied tuple for eventual * freeing. */ if (node->curTuple) heap_freetuple(node->curTuple); node->curTuple = ExecCopySlotHeapTuple(slot); result = heap_getattr(node->curTuple, 1, tdesc, isNull); /* keep scanning subplan to make sure there's only one tuple */ continue; } if (subLinkType == ARRAY_SUBLINK)//数组 { Datum dvalue; bool disnull; found = true; /* stash away current value */ Assert(subplan->firstColType == TupleDescAttr(tdesc, 0)->atttypid); dvalue = slot_getattr(slot, 1, &disnull); astate = accumArrayResultAny(astate, dvalue, disnull, subplan->firstColType, oldcontext); /* keep scanning subplan to collect all values */ continue; } /* cannot allow multiple input tuples for ROWCOMPARE sublink either */ if (subLinkType == ROWCOMPARE_SUBLINK && found)//行比较 ereport(ERROR, (errcode(ERRCODE_CARDINALITY_VIOLATION), errmsg("more than one row returned by a subquery used as an expression"))); found = true;//初始为T /* * For ALL, ANY, and ROWCOMPARE sublinks, load up the Params * representing the columns of the sub-select, and then evaluate the * combining expression. * 对于ALL,ANY和ROWCOMPARE子链接,加载表示子查询列的Params,并解析组合表达式 */ col = 1;//列从1计数 foreach(plst, subplan->paramIds)//循环遍历子查询参数 { int paramid = lfirst_int(plst); ParamExecData *prmdata; prmdata = &(econtext->ecxt_param_exec_vals[paramid]); Assert(prmdata->execPlan == NULL); //获取参数值 prmdata->value = slot_getattr(slot, col, &(prmdata->isnull)); //下一个列 col++; } //解析表达式 rowresult = ExecEvalExprSwitchContext(node->testexpr, econtext, &rownull); if (subLinkType == ANY_SUBLINK) { //ANY : 使用OR语义组合 /* combine across rows per OR semantics */ if (rownull) *isNull = true; else if (DatumGetBool(rowresult)) { result = BoolGetDatum(true); *isNull = false; break; /* needn't look at any more rows */ } } else if (subLinkType == ALL_SUBLINK) { //ALL : 使用AND语义 /* combine across rows per AND semantics */ if (rownull) *isNull = true; else if (!DatumGetBool(rowresult)) { result = BoolGetDatum(false); *isNull = false; break; /* needn't look at any more rows */ } } else { /* must be ROWCOMPARE_SUBLINK */ //这里一定是ROWCOMPARE result = rowresult; *isNull = rownull; } } MemoryContextSwitchTo(oldcontext); if (subLinkType == ARRAY_SUBLINK) { /* We return the result in the caller's context */ //在调用者上下文中返回结果 result = makeArrayResultAny(astate, oldcontext, true); } else if (!found) { /* * deal with empty subplan result. result/isNull were previously * initialized correctly for all sublink types except EXPR and * ROWCOMPARE; for those, return NULL. * subplan没有结果返回,设置result&isNull值 */ if (subLinkType == EXPR_SUBLINK || subLinkType == ROWCOMPARE_SUBLINK) { result = (Datum) 0; *isNull = true; } } //返回结果 return result; } ``` ### 三、跟踪分析 执行SQL： ``` [local]:5432 pg12@testdb=# select * from tbl a where a.id not in (select b.id from t_big_null b); ``` 启动gdb跟踪，设置断点，观察到断点hit 1760次后就会退出，因此设置为忽略前1758次，只跟踪最后2次。 ``` (gdb) info b Num Type Disp Enb Address What 13 breakpoint keep y 0x0000000000721126 in ExecMaterial at nodeMaterial.c:150 breakpoint already hit 1760 times ignore next 3360 hits ... (gdb) b nodeSubplan.c:328 Breakpoint 17 at 0x7303b9: file nodeSubplan.c, line 328. (gdb) del 16 (gdb) info b Num Type Disp Enb Address What 17 breakpoint keep y 0x00000000007303b9 in ExecScanSubPlan at nodeSubplan.c:328 (gdb) ignore 17 1758 Will ignore next 1758 crossings of breakpoint 17. (gdb) c Continuing. ``` 开始跟踪，这是第1759次，这时候从SubPlan获取的数据是id = 10000000 ``` Breakpoint 17, ExecScanSubPlan (node=0x3069268, econtext=0x3068aa0, isNull=0x3068dbd) at nodeSubplan.c:328 328 TupleDesc tdesc = slot->tts_tupleDescriptor; (gdb) n 334 if (subLinkType == EXISTS_SUBLINK) (gdb) 341 if (subLinkType == EXPR_SUBLINK) (gdb) 367 if (subLinkType == ARRAY_SUBLINK) (gdb) 383 if (subLinkType == ROWCOMPARE_SUBLINK && found) (gdb) 388 found = true; (gdb) 395 col = 1; (gdb) 396 foreach(plst, subplan->paramIds) (gdb) 398 int paramid = lfirst_int(plst); (gdb) 401 prmdata = &(econtext->ecxt_param_exec_vals[paramid]); (gdb) 402 Assert(prmdata->execPlan == NULL); (gdb) 403 prmdata->value = slot_getattr(slot, col, &(prmdata->isnull)); (gdb) p *prmdata $109 = {execPlan = 0x0, value = 9999999, isnull = false} (gdb) n 404 col++; (gdb) p *prmdata $110 = {execPlan = 0x0, value = 10000000, isnull = false} (gdb) n 396 foreach(plst, subplan->paramIds) (gdb) ``` 解析表达式 ``` 407 rowresult = ExecEvalExprSwitchContext(node->testexpr, econtext, (gdb) step ExecEvalExprSwitchContext (state=0x3069380, econtext=0x3068aa0, isNull=0x7ffd184750ef) at ../../../src/include/executor/executor.h:306 306 oldContext = MemoryContextSwitchTo(econtext->ecxt_per_tuple_memory); (gdb) n 307 retDatum = state->evalfunc(state, econtext, isNull); (gdb) step ExecInterpExpr (state=0x3069380, econtext=0x3068aa0, isnull=0x7ffd184750ef) at execExprInterp.c:404 404 if (unlikely(state == NULL)) (gdb) n 411 op = state->steps; (gdb) p *state $111 = {tag = {type = T_ExprState}, flags = 6 '\006', resnull = false, resvalue = 0, resultslot = 0x0, steps = 0x3069418, evalfunc = 0x6e2d4d <ExecInterpExpr>, expr = 0x30917a8, evalfunc_private = 0x6e2d4d <ExecInterpExpr>, steps_len = 5, steps_alloc = 16, parent = 0x3068988, ext_params = 0x0, innermost_caseval = 0x0, innermost_casenull = 0x0, innermost_domainval = 0x0, innermost_domainnull = 0x0} (gdb) n 412 resultslot = state->resultslot; (gdb) 413 innerslot = econtext->ecxt_innertuple; (gdb) 414 outerslot = econtext->ecxt_outertuple; (gdb) 415 scanslot = econtext->ecxt_scantuple; (gdb) p *innerslot Cannot access memory at address 0x0 (gdb) p *outerslot Cannot access memory at address 0x0 (gdb) n 418 EEO_DISPATCH(); (gdb) p *scanslot $112 = {type = T_TupleTableSlot, tts_flags = 16, tts_nvalid = 1, tts_ops = 0xc3e780 <TTSOpsBufferHeapTuple>, tts_tupleDescriptor = 0x7fab449c99f0, tts_values = 0x3068bd0, tts_isnull = 0x3068be0, tts_mcxt = 0x3067da0, tts_tid = { ip_blkid = {bi_hi = 0, bi_lo = 0}, ip_posid = 2}, tts_tableOid = 40960} (gdb) p *scanslot->tts_values $113 = 1 (gdb) n 448 CheckOpSlotCompatibility(op, scanslot); (gdb) n 450 slot_getsomeattrs(scanslot, op->d.fetch.last_var); (gdb) 452 EEO_NEXT(); (gdb) 487 int attnum = op->d.var.attnum; (gdb) 491 Assert(attnum >= 0 && attnum < scanslot->tts_nvalid); (gdb) 492 *op->resvalue = scanslot->tts_values[attnum]; (gdb) 493 *op->resnull = scanslot->tts_isnull[attnum]; (gdb) 495 EEO_NEXT(); (gdb) p *op->resvalue $114 = 1 (gdb) n 962 ExecEvalParamExec(state, op, econtext); (gdb) 964 EEO_NEXT(); (gdb) p *op $115 = {opcode = 7224136, resvalue = 0x30698b8, resnull = 0x30698c0, d = {fetch = { last_var = 0, fixed = 23, known_desc = 0x0, kind = 0x0}, var = {attnum = 0, vartype = 23}, wholerow = {var = 0x1700000000, first = false, slow = false, tupdesc = 0x0, junkFilter = 0x0}, assign_var = {resultnum = 0, attnum = 23}, assign_tmp = {resultnum = 0}, constval = {value = 98784247808, isnull = false}, func = {finfo = 0x1700000000, fcinfo_data = 0x0, fn_addr = 0x0, nargs = 0}, boolexpr = {anynull = 0x1700000000, jumpdone = 0}, qualexpr = {jumpdone = 0}, jump = { jumpdone = 0}, nulltest_row = {argdesc = 0x1700000000}, param = {paramid = 0, paramtype = 23}, cparam = {paramfunc = 0x1700000000, paramarg = 0x0, paramid = 0, paramtype = 0}, casetest = {value = 0x1700000000, isnull = 0x0}, make_readonly = { value = 0x1700000000, isnull = 0x0}, iocoerce = {finfo_out = 0x1700000000, fcinfo_data_out = 0x0, finfo_in = 0x0, fcinfo_data_in = 0x0}, sqlvaluefunction = { svf = 0x1700000000}, nextvalueexpr = {seqid = 0, seqtypid = 23}, arrayexpr = { elemvalues = 0x1700000000, elemnulls = 0x0, nelems = 0, elemtype = 0, elemlength = 0, elembyval = false, elemalign = 0 '\000', multidims = false}, arraycoerce = {elemexprstate = 0x1700000000, resultelemtype = 0, amstate = 0x0}, row = {tupdesc = 0x1700000000, elemvalues = 0x0, elemnulls = 0x0}, rowcompare_step = { finfo = 0x1700000000, fcinfo_data = 0x0, fn_addr = 0x0, jumpnull = 0, jumpdone = 0}, rowcompare_final = {rctype = 0}, minmax = {values = 0x1700000000, nulls = 0x0, nelems = 0, op = IS_GREATEST, finfo = 0x0, fcinfo_data = 0x0}, fieldselect = {fieldnum = 0, resulttype = 23, argdesc = 0x0}, fieldstore = { fstore = 0x1700000000, argdesc = 0x0, values = 0x0, nulls = 0x0, ncolumns = 0}, sbsref_subscript = {state = 0x1700000000, off = 0, isupper = false, jumpdone = 0}, sbsref = {state = 0x1700000000}, domaincheck = { constraintname = 0x1700000000 , checkvalue = 0x0, checknull = 0x0, resulttype = 0}, convert_rowtype = { convert = 0x1700000000, indesc = 0x0, outdesc = 0x0, map = 0x0, initialized = false}, scalararrayop = {element_type = 0, useOr = 23, typlen = 0, typbyval = false, typalign = 0 '\000', finfo = 0x0, fcinfo_data = 0x0, fn_addr = 0x0}, xmlexpr = {xexpr = 0x1700000000, named_argvalue = 0x0, named_argnull = 0x0, argvalue = 0x0, argnull = 0x0}, aggref = { astate = 0x1700000000}, grouping_func = {parent = 0x1700000000, clauses = 0x0}, window_func = {wfstate = 0x1700000000}, subplan = {sstate = 0x1700000000}, alternative_subplan = {asstate = 0x1700000000}, agg_deserialize = { aggstate = 0x1700000000, fcinfo_data = 0x0, jumpnull = 0}, agg_strict_input_check = {args = 0x1700000000, nulls = 0x0, nargs = 0, jumpnull = 0}, agg_init_trans = {aggstate = 0x1700000000, pertrans = 0x0, aggcontext = 0x0, setno = 0, transno = 0, setoff = 0, jumpnull = 0}, agg_strict_trans_check = { aggstate = 0x1700000000, setno = 0, transno = 0, setoff = 0, jumpnull = 0}, agg_trans = {aggstate = 0x1700000000, pertrans = 0x0, aggcontext = 0x0, setno = 0, transno = 0, setoff = 0}}} (gdb) p *state $116 = {tag = {type = T_ExprState}, flags = 6 '\006', resnull = false, resvalue = 0, resultslot = 0x0, steps = 0x3069418, evalfunc = 0x6e2d4d <ExecInterpExpr>, expr = 0x30917a8, evalfunc_private = 0x6e2d4d <ExecInterpExpr>, steps_len = 5, steps_alloc = 16, parent = 0x3068988, ext_params = 0x0, innermost_caseval = 0x0, innermost_casenull = 0x0, innermost_domainval = 0x0, innermost_domainnull = 0x0} (gdb) n 634 FunctionCallInfo fcinfo = op->d.func.fcinfo_data; (gdb) 635 NullableDatum *args = fcinfo->args; (gdb) p *fcinfo $117 = {flinfo = 0x3069830, context = 0x0, resultinfo = 0x0, fncollation = 0, isnull = false, nargs = 2, args = 0x30698a8} (gdb) p *fcinfo->args $118 = {value = 1, isnull = false} (gdb) n 640 for (argno = 0; argno < op->d.func.nargs; argno++) (gdb) p op->d.func.nargs $119 = 2 (gdb) p *op $120 = {opcode = 7222440, resvalue = 0x3069388, resnull = 0x3069385, d = {fetch = { last_var = 50763824, fixed = false, known_desc = 0x3069888, kind = 0x96c2b2 <int4eq>}, var = {attnum = 50763824, vartype = 0}, wholerow = { var = 0x3069830, first = 136, slow = 152, tupdesc = 0x96c2b2 <int4eq>, junkFilter = 0x2}, assign_var = {resultnum = 50763824, attnum = 0}, assign_tmp = { resultnum = 50763824}, constval = {value = 50763824, isnull = 136}, func = { finfo = 0x3069830, fcinfo_data = 0x3069888, fn_addr = 0x96c2b2 <int4eq>, nargs = 2}, boolexpr = {anynull = 0x3069830, jumpdone = 50763912}, qualexpr = { jumpdone = 50763824}, jump = {jumpdone = 50763824}, nulltest_row = { argdesc = 0x3069830}, param = {paramid = 50763824, paramtype = 0}, cparam = { paramfunc = 0x3069830, paramarg = 0x3069888, paramid = 9880242, paramtype = 0}, casetest = {value = 0x3069830, isnull = 0x3069888}, make_readonly = { value = 0x3069830, isnull = 0x3069888}, iocoerce = {finfo_out = 0x3069830, fcinfo_data_out = 0x3069888, finfo_in = 0x96c2b2 <int4eq>, fcinfo_data_in = 0x2}, sqlvaluefunction = {svf = 0x3069830}, nextvalueexpr = {seqid = 50763824, seqtypid = 0}, arrayexpr = {elemvalues = 0x3069830, elemnulls = 0x3069888, nelems = 9880242, elemtype = 0, elemlength = 2, elembyval = false, elemalign = 0 '\000', multidims = false}, arraycoerce = {elemexprstate = 0x3069830, resultelemtype = 50763912, amstate = 0x96c2b2 <int4eq>}, row = { tupdesc = 0x3069830, elemvalues = 0x3069888, elemnulls = 0x96c2b2 <int4eq>}, rowcompare_step = {finfo = 0x3069830, fcinfo_data = 0x3069888, fn_addr = 0x96c2b2 <int4eq>, jumpnull = 2, jumpdone = 0}, rowcompare_final = { rctype = 50763824}, minmax = {values = 0x3069830, nulls = 0x3069888, nelems = 9880242, op = IS_GREATEST, finfo = 0x2, fcinfo_data = 0x0}, fieldselect = { fieldnum = -26576, resulttype = 0, argdesc = 0x3069888}, fieldstore = { fstore = 0x3069830, argdesc = 0x3069888, values = 0x96c2b2 <int4eq>, nulls = 0x2, ncolumns = 0}, sbsref_subscript = {state = 0x3069830, off = 50763912, isupper = false, jumpdone = 9880242}, sbsref = {state = 0x3069830}, domaincheck = { constraintname = 0x3069830 "\262\302\226", checkvalue = 0x3069888, checknull = 0x96c2b2 <int4eq>, resulttype = 2}, convert_rowtype = { convert = 0x3069830, indesc = 0x3069888, outdesc = 0x96c2b2 <int4eq>, map = 0x2, initialized = false}, scalararrayop = {element_type = 50763824, useOr = false, typlen = 0, typbyval = 136, typalign = -104 '\230', finfo = 0x96c2b2 <int4eq>, fcinfo_data = 0x2, fn_addr = 0x0}, xmlexpr = {xexpr = 0x3069830, named_argvalue = 0x3069888, named_argnull = 0x96c2b2 <int4eq>, argvalue = 0x2, argnull = 0x0}, aggref = {astate = 0x3069830}, grouping_func = {parent = 0x3069830, clauses = 0x3069888}, window_func = {wfstate = 0x3069830}, subplan = { sstate = 0x3069830}, alternative_subplan = {asstate = 0x3069830}, agg_deserialize = {aggstate = 0x3069830, fcinfo_data = 0x3069888, jumpnull = 9880242}, agg_strict_input_check = {args = 0x3069830, nulls = 0x3069888, nargs = 9880242, jumpnull = 0}, agg_init_trans = {aggstate = 0x3069830, pertrans = 0x3069888, aggcontext = 0x96c2b2 <int4eq>, setno = 2, transno = 0, setoff = 0, jumpnull = 0}, agg_strict_trans_check = {aggstate = 0x3069830, setno = 50763912, transno = 0, setoff = 9880242, jumpnull = 0}, agg_trans = { aggstate = 0x3069830, pertrans = 0x3069888, aggcontext = 0x96c2b2 <int4eq>, setno = 2, transno = 0, setoff = 0}}} (gdb) p op->d->func $121 = {finfo = 0x3069830, fcinfo_data = 0x3069888, fn_addr = 0x96c2b2 <int4eq>, nargs = 2} (gdb) p op->d->func->finfo $122 = (FmgrInfo *) 0x3069830 (gdb) p *op->d->func->finfo $123 = {fn_addr = 0x96c2b2 <int4eq>, fn_oid = 65, fn_nargs = 2, fn_strict = true, fn_retset = false, fn_stats = 2 '\002', fn_extra = 0x0, fn_mcxt = 0x3067da0, fn_expr = 0x30917a8} (gdb) p *op->d->func->fcinfo_data $124 = {flinfo = 0x3069830, context = 0x0, resultinfo = 0x0, fncollation = 0, isnull = false, nargs = 2, args = 0x30698a8} (gdb) p *op->d->func->fcinfo_data->flinfo $125 = {fn_addr = 0x96c2b2 <int4eq>, fn_oid = 65, fn_nargs = 2, fn_strict = true, fn_retset = false, fn_stats = 2 '\002', fn_extra = 0x0, fn_mcxt = 0x3067da0, fn_expr = 0x30917a8} (gdb) p *op->d->func->fcinfo_data->args $126 = {value = 1, isnull = false} (gdb) n 642 if (args[argno].isnull) (gdb) 640 for (argno = 0; argno < op->d.func.nargs; argno++) (gdb) 642 if (args[argno].isnull) (gdb) 640 for (argno = 0; argno < op->d.func.nargs; argno++) (gdb) 648 fcinfo->isnull = false; (gdb) p *args $127 = {value = 1, isnull = false} (gdb) n 649 d = op->d.func.fn_addr(fcinfo); (gdb) 650 *op->resvalue = d; (gdb) p d $128 = 0 (gdb) n 651 *op->resnull = fcinfo->isnull; (gdb) 654 EEO_NEXT(); (gdb) 425 goto out; (gdb) n 1747 *isnull = state->resnull; (gdb) 1748 return state->resvalue; (gdb) p *state $129 = {tag = {type = T_ExprState}, flags = 6 '\006', resnull = false, resvalue = 0, resultslot = 0x0, steps = 0x3069418, evalfunc = 0x6e2d4d <ExecInterpExpr>, expr = 0x30917a8, evalfunc_private = 0x6e2d4d <ExecInterpExpr>, steps_len = 5, steps_alloc = 16, parent = 0x3068988, ext_params = 0x0, innermost_caseval = 0x0, innermost_casenull = 0x0, innermost_domainval = 0x0, innermost_domainnull = 0x0} (gdb) n 1749 } (gdb) ExecEvalExprSwitchContext (state=0x3069380, econtext=0x3068aa0, isNull=0x7ffd184750ef) at ../../../src/include/executor/executor.h:308 308 MemoryContextSwitchTo(oldContext); (gdb) 309 return retDatum; (gdb) p retDatum $130 = 0 (gdb) n 310 } (gdb) ``` 这是第1760次调用 ``` ExecScanSubPlan (node=0x3069268, econtext=0x3068aa0, isNull=0x3068dbd) at nodeSubplan.c:410 410 if (subLinkType == ANY_SUBLINK) (gdb) 413 if (rownull) (gdb) 415 else if (DatumGetBool(rowresult)) (gdb) p rowresult $131 = 0 (gdb) n 326 slot = ExecProcNode(planstate)) (gdb) 324 for (slot = ExecProcNode(planstate); (gdb) 325 !TupIsNull(slot); (gdb) Breakpoint 17, ExecScanSubPlan (node=0x3069268, econtext=0x3068aa0, isNull=0x3068dbd) at nodeSubplan.c:328 328 TupleDesc tdesc = slot->tts_tupleDescriptor; (gdb) 334 if (subLinkType == EXISTS_SUBLINK) (gdb) 341 if (subLinkType == EXPR_SUBLINK) (gdb) 367 if (subLinkType == ARRAY_SUBLINK) (gdb) 383 if (subLinkType == ROWCOMPARE_SUBLINK && found) (gdb) 388 found = true; (gdb) 395 col = 1; (gdb) p *slot->tts_values $132 = 10000000 --> 上一次的数据 (gdb) n 396 foreach(plst, subplan->paramIds) (gdb) 398 int paramid = lfirst_int(plst); (gdb) 401 prmdata = &(econtext->ecxt_param_exec_vals[paramid]); (gdb) 402 Assert(prmdata->execPlan == NULL); (gdb) p *prmdata $133 = {execPlan = 0x0, value = 10000000, isnull = false} (gdb) n 403 prmdata->value = slot_getattr(slot, col, &(prmdata->isnull)); (gdb) 404 col++; (gdb) p *prmdata $134 = {execPlan = 0x0, value = 1, isnull = false} --> 本次数据，值为1 (gdb) info b Num Type Disp Enb Address What 17 breakpoint keep y 0x00000000007303b9 in ExecScanSubPlan at nodeSubplan.c:328 breakpoint already hit 1760 times (gdb) n 396 foreach(plst, subplan->paramIds) (gdb) 407 rowresult = ExecEvalExprSwitchContext(node->testexpr, econtext, (gdb) 410 if (subLinkType == ANY_SUBLINK) (gdb) 413 if (rownull) (gdb) 415 else if (DatumGetBool(rowresult)) (gdb) 417 result = BoolGetDatum(true); (gdb) 418 *isNull = false; (gdb) 419 break; /* needn't look at any more rows */ (gdb) 442 MemoryContextSwitchTo(oldcontext); (gdb) 444 if (subLinkType == ARRAY_SUBLINK) (gdb) 449 else if (!found) (gdb) 464 return result; (gdb) (gdb) p result $135 = 1 --> 满足条件 ``` **DONE** ### 四、参考资料 N/A