/*------------------------------------------------------------------------- * * deparse.c * Query deparser for treasuredata_fdw * * Portions Copyright (c) 2016, Mitsunori Komatsu * * IDENTIFICATION * deparse.c * *------------------------------------------------------------------------- */ #include "postgres.h" #include "treasuredata_fdw.h" #include "access/heapam.h" #include "access/htup_details.h" #include "access/sysattr.h" #include "access/transam.h" #include "catalog/pg_collation.h" #include "catalog/pg_namespace.h" #include "catalog/pg_operator.h" #include "catalog/pg_proc.h" #include "catalog/pg_type.h" #include "commands/defrem.h" #include "nodes/nodeFuncs.h" #include "optimizer/clauses.h" #include "optimizer/var.h" #include "parser/parsetree.h" #include "utils/builtins.h" #include "utils/lsyscache.h" #include "utils/rel.h" #include "utils/syscache.h" /* * Global context for foreign_expr_walker's search of an expression tree. */ typedef struct foreign_glob_cxt { PlannerInfo *root; /* global planner state */ RelOptInfo *foreignrel; /* the foreign relation we are planning for */ } foreign_glob_cxt; /* * Local (per-tree-level) context for foreign_expr_walker's search. * This is concerned with identifying collations used in the expression. */ typedef enum { FDW_COLLATE_NONE, /* expression is of a noncollatable type, or * it has default collation that is not * traceable to a foreign Var */ FDW_COLLATE_SAFE, /* collation derives from a foreign Var */ FDW_COLLATE_UNSAFE /* collation is non-default and derives from * something other than a foreign Var */ } FDWCollateState; typedef struct foreign_loc_cxt { Oid collation; /* OID of current collation, if any */ FDWCollateState state; /* state of current collation choice */ } foreign_loc_cxt; /* * Context for deparseExpr */ typedef struct deparse_expr_cxt { PlannerInfo *root; /* global planner state */ RelOptInfo *foreignrel; /* the foreign relation we are planning for */ QueryEngineType query_engine_type; /* Query engine type */ StringInfo buf; /* output buffer to append to */ List **params_list; /* exprs that will become remote Params */ } deparse_expr_cxt; /* * Functions to determine whether an expression can be evaluated safely on * remote server. */ static bool foreign_expr_walker(Node *node, foreign_glob_cxt *glob_cxt, foreign_loc_cxt *outer_cxt); static bool is_builtin(Oid procid); /* * Functions to construct string representation of a node tree. */ static void deparseTargetList(StringInfo buf, PlannerInfo *root, Index rtindex, Relation rel, bool is_returning, Bitmapset *attrs_used, List **retrieved_attrs, QueryEngineType query_engine_type); static void deparseReturningList(StringInfo buf, PlannerInfo *root, Index rtindex, Relation rel, bool trig_after_row, List *returningList, List **retrieved_attrs); static void deparseColumnRef(StringInfo buf, int varno, int varattno, PlannerInfo *root, QueryEngineType query_engine_type); static void deparseRelation(StringInfo buf, Relation rel, QueryEngineType query_engine_type); static void deparseExpr(Expr *expr, deparse_expr_cxt *context); static void deparseVar(Var *node, deparse_expr_cxt *context); static void deparseConst(Const *node, deparse_expr_cxt *context); static void deparseParam(Param *node, deparse_expr_cxt *context); static void deparseArrayRef(ArrayRef *node, deparse_expr_cxt *context); static void deparseFuncExpr(FuncExpr *node, deparse_expr_cxt *context); static void deparseOpExpr(OpExpr *node, deparse_expr_cxt *context); static void deparseOperatorName(StringInfo buf, Form_pg_operator opform); static void deparseDistinctExpr(DistinctExpr *node, deparse_expr_cxt *context); static void deparseScalarArrayOpExpr(ScalarArrayOpExpr *node, deparse_expr_cxt *context); static void deparseRelabelType(RelabelType *node, deparse_expr_cxt *context); static void deparseBoolExpr(BoolExpr *node, deparse_expr_cxt *context); static void deparseNullTest(NullTest *node, deparse_expr_cxt *context); static void deparseArrayExpr(ArrayExpr *node, deparse_expr_cxt *context); static void printRemoteParam(int paramindex, Oid paramtype, int32 paramtypmod, deparse_expr_cxt *context); static void printRemotePlaceholder(Oid paramtype, int32 paramtypmod, deparse_expr_cxt *context); static bool isAllowedFunction(FuncExpr *node); static char *td_quote_identifier(const char *s, QueryEngineType query_engine_type); static void td_deparse_string(StringInfo buf, const char *val, bool is_str, QueryEngineType query_engine_type); static void deparseStringLiteralForQueryEngineType(StringInfo buf, const char *val, QueryEngineType query_engine_type); /* * Examine each qual clause in input_conds, and classify them into two groups, * which are returned as two lists: * - remote_conds contains expressions that can be evaluated remotely * - local_conds contains expressions that can't be evaluated remotely */ void classifyConditions(PlannerInfo *root, RelOptInfo *baserel, List *input_conds, List **remote_conds, List **local_conds) { ListCell *lc; *remote_conds = NIL; *local_conds = NIL; foreach(lc, input_conds) { RestrictInfo *ri = (RestrictInfo *) lfirst(lc); if (is_foreign_expr(root, baserel, ri->clause)) *remote_conds = lappend(*remote_conds, ri); else *local_conds = lappend(*local_conds, ri); } } /* * Returns true if given expr is safe to evaluate on the foreign server. */ bool is_foreign_expr(PlannerInfo *root, RelOptInfo *baserel, Expr *expr) { foreign_glob_cxt glob_cxt; foreign_loc_cxt loc_cxt; /* * Check that the expression consists of nodes that are safe to execute * remotely. */ glob_cxt.root = root; glob_cxt.foreignrel = baserel; loc_cxt.collation = InvalidOid; loc_cxt.state = FDW_COLLATE_NONE; if (!foreign_expr_walker((Node *) expr, &glob_cxt, &loc_cxt)) return false; /* Expressions examined here should be boolean, ie noncollatable */ Assert(loc_cxt.collation == InvalidOid); Assert(loc_cxt.state == FDW_COLLATE_NONE); /* * An expression which includes any mutable functions can't be sent over * because its result is not stable. For example, sending now() remote * side could cause confusion from clock offsets. Future versions might * be able to make this choice with more granularity. (We check this last * because it requires a lot of expensive catalog lookups.) */ if (contain_mutable_functions((Node *) expr)) return false; /* OK to evaluate on the remote server */ return true; } /* * Check if expression is safe to execute remotely, and return true if so. * * In addition, *outer_cxt is updated with collation information. * * We must check that the expression contains only node types we can deparse, * that all types/functions/operators are safe to send (which we approximate * as being built-in), and that all collations used in the expression derive * from Vars of the foreign table. Because of the latter, the logic is * pretty close to assign_collations_walker() in parse_collate.c, though we * can assume here that the given expression is valid. */ static bool foreign_expr_walker(Node *node, foreign_glob_cxt *glob_cxt, foreign_loc_cxt *outer_cxt) { bool check_type = true; foreign_loc_cxt inner_cxt; Oid collation; FDWCollateState state; /* Need do nothing for empty subexpressions */ if (node == NULL) return true; /* Set up inner_cxt for possible recursion to child nodes */ inner_cxt.collation = InvalidOid; inner_cxt.state = FDW_COLLATE_NONE; switch (nodeTag(node)) { case T_Var: { Var *var = (Var *) node; /* * If the Var is from the foreign table, we consider its * collation (if any) safe to use. If it is from another * table, we treat its collation the same way as we would a * Param's collation, ie it's not safe for it to have a * non-default collation. */ if (var->varno == glob_cxt->foreignrel->relid && var->varlevelsup == 0) { /* Var belongs to foreign table */ /* * System columns other than ctid should not be sent to * the remote, since we don't make any effort to ensure * that local and remote values match (tableoid, in * particular, almost certainly doesn't match). */ if (var->varattno < 0 && var->varattno != SelfItemPointerAttributeNumber) return false; /* Else check the collation */ collation = var->varcollid; state = OidIsValid(collation) ? FDW_COLLATE_SAFE : FDW_COLLATE_NONE; } else { /* Var belongs to some other table */ collation = var->varcollid; if (collation == InvalidOid || collation == DEFAULT_COLLATION_OID) { /* * It's noncollatable, or it's safe to combine with a * collatable foreign Var, so set state to NONE. */ state = FDW_COLLATE_NONE; } else { /* * Do not fail right away, since the Var might appear * in a collation-insensitive context. */ state = FDW_COLLATE_UNSAFE; } } } break; case T_Const: { Const *c = (Const *) node; /* * If the constant has nondefault collation, either it's of a * non-builtin type, or it reflects folding of a CollateExpr. * It's unsafe to send to the remote unless it's used in a * non-collation-sensitive context. */ collation = c->constcollid; if (collation == InvalidOid || collation == DEFAULT_COLLATION_OID) state = FDW_COLLATE_NONE; else state = FDW_COLLATE_UNSAFE; } break; case T_Param: /* This FDW doesn't support param push down */ return false; case T_ArrayRef: { ArrayRef *ar = (ArrayRef *) node; /* Assignment should not be in restrictions. */ if (ar->refassgnexpr != NULL) return false; /* * Recurse to remaining subexpressions. Since the array * subscripts must yield (noncollatable) integers, they won't * affect the inner_cxt state. */ if (!foreign_expr_walker((Node *) ar->refupperindexpr, glob_cxt, &inner_cxt)) return false; if (!foreign_expr_walker((Node *) ar->reflowerindexpr, glob_cxt, &inner_cxt)) return false; if (!foreign_expr_walker((Node *) ar->refexpr, glob_cxt, &inner_cxt)) return false; /* * Array subscripting should yield same collation as input, * but for safety use same logic as for function nodes. */ collation = ar->refcollid; if (collation == InvalidOid) state = FDW_COLLATE_NONE; else if (inner_cxt.state == FDW_COLLATE_SAFE && collation == inner_cxt.collation) state = FDW_COLLATE_SAFE; else if (collation == DEFAULT_COLLATION_OID) state = FDW_COLLATE_NONE; else state = FDW_COLLATE_UNSAFE; } break; case T_FuncExpr: { FuncExpr *fe = (FuncExpr *) node; /* * If function used by the expression is not built-in, it * can't be sent to remote because it might have incompatible * semantics on remote side. */ if (!is_builtin(fe->funcid)) return false; /* * Recurse to input subexpressions. */ if (!foreign_expr_walker((Node *) fe->args, glob_cxt, &inner_cxt)) return false; /* * If function's input collation is not derived from a foreign * Var, it can't be sent to remote. */ if (fe->inputcollid == InvalidOid) /* OK, inputs are all noncollatable */ ; else if (inner_cxt.state != FDW_COLLATE_SAFE || fe->inputcollid != inner_cxt.collation) return false; if (!isAllowedFunction(fe)) return false; /* * Detect whether node is introducing a collation not derived * from a foreign Var. (If so, we just mark it unsafe for now * rather than immediately returning false, since the parent * node might not care.) */ collation = fe->funccollid; if (collation == InvalidOid) state = FDW_COLLATE_NONE; else if (inner_cxt.state == FDW_COLLATE_SAFE && collation == inner_cxt.collation) state = FDW_COLLATE_SAFE; else if (collation == DEFAULT_COLLATION_OID) state = FDW_COLLATE_NONE; else state = FDW_COLLATE_UNSAFE; } break; case T_OpExpr: case T_DistinctExpr: /* struct-equivalent to OpExpr */ { OpExpr *oe = (OpExpr *) node; /* * Similarly, only built-in operators can be sent to remote. * (If the operator is, surely its underlying function is * too.) */ if (!is_builtin(oe->opno)) return false; /* * Recurse to input subexpressions. */ if (!foreign_expr_walker((Node *) oe->args, glob_cxt, &inner_cxt)) return false; /* * If operator's input collation is not derived from a foreign * Var, it can't be sent to remote. */ if (oe->inputcollid == InvalidOid) /* OK, inputs are all noncollatable */ ; else if (inner_cxt.state != FDW_COLLATE_SAFE || oe->inputcollid != inner_cxt.collation) return false; /* Result-collation handling is same as for functions */ collation = oe->opcollid; if (collation == InvalidOid) state = FDW_COLLATE_NONE; else if (inner_cxt.state == FDW_COLLATE_SAFE && collation == inner_cxt.collation) state = FDW_COLLATE_SAFE; else if (collation == DEFAULT_COLLATION_OID) state = FDW_COLLATE_NONE; else state = FDW_COLLATE_UNSAFE; } break; case T_ScalarArrayOpExpr: { ScalarArrayOpExpr *oe = (ScalarArrayOpExpr *) node; /* * Again, only built-in operators can be sent to remote. */ if (!is_builtin(oe->opno)) return false; /* * Recurse to input subexpressions. */ if (!foreign_expr_walker((Node *) oe->args, glob_cxt, &inner_cxt)) return false; /* * If operator's input collation is not derived from a foreign * Var, it can't be sent to remote. */ if (oe->inputcollid == InvalidOid) /* OK, inputs are all noncollatable */ ; else if (inner_cxt.state != FDW_COLLATE_SAFE || oe->inputcollid != inner_cxt.collation) return false; { char *opname; Form_pg_operator form; /* Retrieve information about the operator from system catalog. */ HeapTuple tuple = SearchSysCache1(OPEROID, ObjectIdGetDatum(oe->opno)); if (!HeapTupleIsValid(tuple)) elog(ERROR, "cache lookup failed for operator %u", oe->opno); form = (Form_pg_operator) GETSTRUCT(tuple); opname = NameStr(form->oprname); /* * x IN (a, b, c) => opname:'=', node->useOr:true (Available on TD) * x NOT IN (a, b, c) => opname:'<>', node->useOr:false (Available on TD) * x = ANY('{a,b,c}') => opname:'=', node->useOr:true (Can be converted to `IN`) * * Other than the above are not supported */ if (!(strcmp(opname, "=") == 0 && oe->useOr) && !(strcmp(opname, "<>") == 0 && !oe->useOr)) { elog(DEBUG1, "Unsupported operator combination in scalar array. opname=[%s], node->useOr=%d", opname, oe->useOr); ReleaseSysCache(tuple); return false; } ReleaseSysCache(tuple); } /* Output is always boolean and so noncollatable. */ collation = InvalidOid; state = FDW_COLLATE_NONE; } break; case T_RelabelType: { RelabelType *r = (RelabelType *) node; /* * Recurse to input subexpression. */ if (!foreign_expr_walker((Node *) r->arg, glob_cxt, &inner_cxt)) return false; /* * RelabelType must not introduce a collation not derived from * an input foreign Var (same logic as for a real function). */ collation = r->resultcollid; if (collation == InvalidOid) state = FDW_COLLATE_NONE; else if (inner_cxt.state == FDW_COLLATE_SAFE && collation == inner_cxt.collation) state = FDW_COLLATE_SAFE; else if (collation == DEFAULT_COLLATION_OID) state = FDW_COLLATE_NONE; else state = FDW_COLLATE_UNSAFE; } break; case T_BoolExpr: { BoolExpr *b = (BoolExpr *) node; /* * Recurse to input subexpressions. */ if (!foreign_expr_walker((Node *) b->args, glob_cxt, &inner_cxt)) return false; /* Output is always boolean and so noncollatable. */ collation = InvalidOid; state = FDW_COLLATE_NONE; } break; case T_NullTest: { NullTest *nt = (NullTest *) node; /* * Recurse to input subexpressions. */ if (!foreign_expr_walker((Node *) nt->arg, glob_cxt, &inner_cxt)) return false; /* Output is always boolean and so noncollatable. */ collation = InvalidOid; state = FDW_COLLATE_NONE; } break; case T_ArrayExpr: { ArrayExpr *a = (ArrayExpr *) node; /* * Recurse to input subexpressions. */ if (!foreign_expr_walker((Node *) a->elements, glob_cxt, &inner_cxt)) return false; /* * ArrayExpr must not introduce a collation not derived from * an input foreign Var (same logic as for a function). */ collation = a->array_collid; if (collation == InvalidOid) state = FDW_COLLATE_NONE; else if (inner_cxt.state == FDW_COLLATE_SAFE && collation == inner_cxt.collation) state = FDW_COLLATE_SAFE; else if (collation == DEFAULT_COLLATION_OID) state = FDW_COLLATE_NONE; else state = FDW_COLLATE_UNSAFE; } break; case T_List: { List *l = (List *) node; ListCell *lc; /* * Recurse to component subexpressions. */ foreach(lc, l) { if (!foreign_expr_walker((Node *) lfirst(lc), glob_cxt, &inner_cxt)) return false; } /* * When processing a list, collation state just bubbles up * from the list elements. */ collation = inner_cxt.collation; state = inner_cxt.state; /* Don't apply exprType() to the list. */ check_type = false; } break; default: /* * If it's anything else, assume it's unsafe. This list can be * expanded later, but don't forget to add deparse support below. */ return false; } /* * If result type of given expression is not built-in, it can't be sent to * remote because it might have incompatible semantics on remote side. */ if (check_type && !is_builtin(exprType(node))) return false; /* * Now, merge my collation information into my parent's state. */ if (state > outer_cxt->state) { /* Override previous parent state */ outer_cxt->collation = collation; outer_cxt->state = state; } else if (state == outer_cxt->state) { /* Merge, or detect error if there's a collation conflict */ switch (state) { case FDW_COLLATE_NONE: /* Nothing + nothing is still nothing */ break; case FDW_COLLATE_SAFE: if (collation != outer_cxt->collation) { /* * Non-default collation always beats default. */ if (outer_cxt->collation == DEFAULT_COLLATION_OID) { /* Override previous parent state */ outer_cxt->collation = collation; } else if (collation != DEFAULT_COLLATION_OID) { /* * Conflict; show state as indeterminate. We don't * want to "return false" right away, since parent * node might not care about collation. */ outer_cxt->state = FDW_COLLATE_UNSAFE; } } break; case FDW_COLLATE_UNSAFE: /* We're still conflicted ... */ break; } } /* It looks OK */ return true; } /* * Return true if given object is one of PostgreSQL's built-in objects. * * We use FirstBootstrapObjectId as the cutoff, so that we only consider * objects with hand-assigned OIDs to be "built in", not for instance any * function or type defined in the information_schema. * * Our constraints for dealing with types are tighter than they are for * functions or operators: we want to accept only types that are in pg_catalog, * else format_type might incorrectly fail to schema-qualify their names. * (This could be fixed with some changes to format_type, but for now there's * no need.) Thus we must exclude information_schema types. * * XXX there is a problem with this, which is that the set of built-in * objects expands over time. Something that is built-in to us might not * be known to the remote server, if it's of an older version. But keeping * track of that would be a huge exercise. */ static bool is_builtin(Oid oid) { return (oid < FirstBootstrapObjectId); } /* * Construct a simple SELECT statement that retrieves desired columns * of the specified foreign table, and append it to "buf". The output * contains just "SELECT ... FROM tablename". * * We also create an integer List of the columns being retrieved, which is * returned to *retrieved_attrs. */ void deparseSelectSql(StringInfo buf, PlannerInfo *root, RelOptInfo *baserel, Bitmapset *attrs_used, List **retrieved_attrs, QueryEngineType query_engine_type) { RangeTblEntry *rte = planner_rt_fetch(baserel->relid, root); Relation rel; /* * Core code already has some lock on each rel being planned, so we can * use NoLock here. */ rel = heap_open(rte->relid, NoLock); /* * Construct SELECT list */ appendStringInfoString(buf, "SELECT "); deparseTargetList(buf, root, baserel->relid, rel, false, attrs_used, retrieved_attrs, query_engine_type); /* * Construct FROM clause */ appendStringInfoString(buf, " FROM "); deparseRelation(buf, rel, query_engine_type); heap_close(rel, NoLock); } /* * Emit a target list that retrieves the columns specified in attrs_used. * This is used for both SELECT and RETURNING targetlists; the is_returning * parameter is true only for a RETURNING targetlist. * * The tlist text is appended to buf, and we also create an integer List * of the columns being retrieved, which is returned to *retrieved_attrs. */ static void deparseTargetList(StringInfo buf, PlannerInfo *root, Index rtindex, Relation rel, bool is_returning, Bitmapset *attrs_used, List **retrieved_attrs, QueryEngineType query_engine_type) { TupleDesc tupdesc = RelationGetDescr(rel); bool have_wholerow; bool first; int i; *retrieved_attrs = NIL; /* If there's a whole-row reference, we'll need all the columns. */ have_wholerow = bms_is_member(0 - FirstLowInvalidHeapAttributeNumber, attrs_used); first = true; for (i = 1; i <= tupdesc->natts; i++) { Form_pg_attribute attr = tupdesc->attrs[i - 1]; /* Ignore dropped attributes. */ if (attr->attisdropped) continue; if (have_wholerow || bms_is_member(i - FirstLowInvalidHeapAttributeNumber, attrs_used)) { if (!first) appendStringInfoString(buf, ", "); else if (is_returning) appendStringInfoString(buf, " RETURNING "); first = false; deparseColumnRef(buf, rtindex, i, root, query_engine_type); *retrieved_attrs = lappend_int(*retrieved_attrs, i); } } /* * Add ctid if needed. We currently don't support retrieving any other * system columns. */ if (bms_is_member(SelfItemPointerAttributeNumber - FirstLowInvalidHeapAttributeNumber, attrs_used)) { if (!first) appendStringInfoString(buf, ", "); else if (is_returning) appendStringInfoString(buf, " RETURNING "); first = false; appendStringInfoString(buf, "ctid"); *retrieved_attrs = lappend_int(*retrieved_attrs, SelfItemPointerAttributeNumber); } /* Don't generate bad syntax if no undropped columns */ if (first && !is_returning) appendStringInfoString(buf, "CAST(NULL AS TINYINT)"); } /* * Deparse WHERE clauses in given list of RestrictInfos and append them to buf. * * baserel is the foreign table we're planning for. * * If no WHERE clause already exists in the buffer, is_first should be true. * * If params is not NULL, it receives a list of Params and other-relation Vars * used in the clauses; these values must be transmitted to the remote server * as parameter values. * * If params is NULL, we're generating the query for EXPLAIN purposes, * so Params and other-relation Vars should be replaced by dummy values. */ void appendWhereClause(StringInfo buf, PlannerInfo *root, RelOptInfo *baserel, List *exprs, bool is_first, List **params, QueryEngineType query_engine_type) { deparse_expr_cxt context; ListCell *lc; if (params) *params = NIL; /* initialize result list to empty */ /* Set up context struct for recursion */ context.root = root; context.foreignrel = baserel; context.query_engine_type = query_engine_type; context.buf = buf; context.params_list = params; foreach(lc, exprs) { RestrictInfo *ri = (RestrictInfo *) lfirst(lc); /* Connect expressions with "AND" and parenthesize each condition. */ if (is_first) appendStringInfoString(buf, " WHERE "); else appendStringInfoString(buf, " AND "); appendStringInfoChar(buf, '('); deparseExpr(ri->clause, &context); appendStringInfoChar(buf, ')'); is_first = false; } } /* * deparse remote INSERT statement * * The statement text is appended to buf, and we also create an integer List * of the columns being retrieved by RETURNING (if any), which is returned * to *retrieved_attrs. */ void deparseInsertSql(StringInfo buf, PlannerInfo *root, Index rtindex, Relation rel, List *targetAttrs, bool doNothing, List *returningList, List **retrieved_attrs) { AttrNumber pindex; bool first; ListCell *lc; appendStringInfoString(buf, "INSERT INTO "); /* TODO: Support Hive */ deparseRelation(buf, rel, QUERY_ENGINE_PRESTO); if (targetAttrs) { appendStringInfoChar(buf, '('); first = true; foreach(lc, targetAttrs) { int attnum = lfirst_int(lc); if (!first) appendStringInfoString(buf, ", "); first = false; /* TODO: Support Hive */ deparseColumnRef(buf, rtindex, attnum, root, QUERY_ENGINE_PRESTO); } appendStringInfoString(buf, ") VALUES ("); pindex = 1; first = true; foreach(lc, targetAttrs) { if (!first) appendStringInfoString(buf, ", "); first = false; appendStringInfo(buf, "$%d", pindex); pindex++; } appendStringInfoChar(buf, ')'); } else appendStringInfoString(buf, " DEFAULT VALUES"); if (doNothing) appendStringInfoString(buf, " ON CONFLICT DO NOTHING"); deparseReturningList(buf, root, rtindex, rel, rel->trigdesc && rel->trigdesc->trig_insert_after_row, returningList, retrieved_attrs); } /* * deparse remote UPDATE statement * * The statement text is appended to buf, and we also create an integer List * of the columns being retrieved by RETURNING (if any), which is returned * to *retrieved_attrs. */ void deparseUpdateSql(StringInfo buf, PlannerInfo *root, Index rtindex, Relation rel, List *targetAttrs, List *returningList, List **retrieved_attrs) { AttrNumber pindex; bool first; ListCell *lc; appendStringInfoString(buf, "UPDATE "); /* TODO: Support Hive */ deparseRelation(buf, rel, QUERY_ENGINE_PRESTO); appendStringInfoString(buf, " SET "); pindex = 2; /* ctid is always the first param */ first = true; foreach(lc, targetAttrs) { int attnum = lfirst_int(lc); if (!first) appendStringInfoString(buf, ", "); first = false; /* TODO: Support Hive */ deparseColumnRef(buf, rtindex, attnum, root, QUERY_ENGINE_PRESTO); appendStringInfo(buf, " = $%d", pindex); pindex++; } appendStringInfoString(buf, " WHERE ctid = $1"); deparseReturningList(buf, root, rtindex, rel, rel->trigdesc && rel->trigdesc->trig_update_after_row, returningList, retrieved_attrs); } /* * deparse remote DELETE statement * * The statement text is appended to buf, and we also create an integer List * of the columns being retrieved by RETURNING (if any), which is returned * to *retrieved_attrs. */ void deparseDeleteSql(StringInfo buf, PlannerInfo *root, Index rtindex, Relation rel, List *returningList, List **retrieved_attrs) { appendStringInfoString(buf, "DELETE FROM "); /* TODO: Support Hive */ deparseRelation(buf, rel, QUERY_ENGINE_PRESTO); appendStringInfoString(buf, " WHERE ctid = $1"); deparseReturningList(buf, root, rtindex, rel, rel->trigdesc && rel->trigdesc->trig_delete_after_row, returningList, retrieved_attrs); } /* * Add a RETURNING clause, if needed, to an INSERT/UPDATE/DELETE. */ static void deparseReturningList(StringInfo buf, PlannerInfo *root, Index rtindex, Relation rel, bool trig_after_row, List *returningList, List **retrieved_attrs) { Bitmapset *attrs_used = NULL; if (trig_after_row) { /* whole-row reference acquires all non-system columns */ attrs_used = bms_make_singleton(0 - FirstLowInvalidHeapAttributeNumber); } if (returningList != NIL) { /* * We need the attrs, non-system and system, mentioned in the local * query's RETURNING list. */ pull_varattnos((Node *) returningList, rtindex, &attrs_used); } if (attrs_used != NULL) /* TODO: Support Hive */ deparseTargetList(buf, root, rtindex, rel, true, attrs_used, retrieved_attrs, QUERY_ENGINE_PRESTO); else *retrieved_attrs = NIL; } /* * Construct SELECT statement to acquire size in blocks of given relation. * * Note: we use local definition of block size, not remote definition. * This is perhaps debatable. * * Note: pg_relation_size() exists in 8.1 and later. */ void deparseAnalyzeSizeSql(StringInfo buf, Relation rel) { StringInfoData relname; /* We'll need the remote relation name as a literal. */ initStringInfo(&relname); /* TODO: Support Hive */ deparseRelation(&relname, rel, QUERY_ENGINE_PRESTO); appendStringInfoString(buf, "SELECT pg_catalog.pg_relation_size("); deparseStringLiteral(buf, relname.data); appendStringInfo(buf, "::pg_catalog.regclass) / %d", BLCKSZ); } /* * Construct SELECT statement to acquire sample rows of given relation. * * SELECT command is appended to buf, and list of columns retrieved * is returned to *retrieved_attrs. */ void deparseAnalyzeSql(StringInfo buf, Relation rel, List **retrieved_attrs) { Oid relid = RelationGetRelid(rel); TupleDesc tupdesc = RelationGetDescr(rel); int i; char *colname; List *options; ListCell *lc; bool first = true; *retrieved_attrs = NIL; appendStringInfoString(buf, "SELECT "); for (i = 0; i < tupdesc->natts; i++) { /* Ignore dropped columns. */ if (tupdesc->attrs[i]->attisdropped) continue; if (!first) appendStringInfoString(buf, ", "); first = false; /* Use attribute name or column_name option. */ colname = NameStr(tupdesc->attrs[i]->attname); options = GetForeignColumnOptions(relid, i + 1); foreach(lc, options) { DefElem *def = (DefElem *) lfirst(lc); if (strcmp(def->defname, "column_name") == 0) { colname = defGetString(def); break; } } appendStringInfoString(buf, quote_identifier(colname)); *retrieved_attrs = lappend_int(*retrieved_attrs, i + 1); } /* Don't generate bad syntax for zero-column relation. */ if (first) appendStringInfoString(buf, "NULL"); /* * Construct FROM clause */ appendStringInfoString(buf, " FROM "); /* TODO: Support Hive */ deparseRelation(buf, rel, QUERY_ENGINE_PRESTO); } /* * Construct name to use for given column, and emit it into buf. * If it has a column_name FDW option, use that instead of attribute name. */ static void deparseColumnRef(StringInfo buf, int varno, int varattno, PlannerInfo *root, QueryEngineType query_engine_type) { RangeTblEntry *rte; char *colname = NULL; List *options; ListCell *lc; /* varno must not be any of OUTER_VAR, INNER_VAR and INDEX_VAR. */ Assert(!IS_SPECIAL_VARNO(varno)); /* Get RangeTblEntry from array in PlannerInfo. */ rte = planner_rt_fetch(varno, root); /* * If it's a column of a foreign table, and it has the column_name FDW * option, use that value. */ options = GetForeignColumnOptions(rte->relid, varattno); foreach(lc, options) { DefElem *def = (DefElem *) lfirst(lc); if (strcmp(def->defname, "column_name") == 0) { colname = defGetString(def); break; } } /* * If it's a column of a regular table or it doesn't have column_name FDW * option, use attribute name. */ if (colname == NULL) colname = get_relid_attribute_name(rte->relid, varattno); appendStringInfoString(buf, td_quote_identifier(colname, query_engine_type)); } /* * Append remote name of specified foreign table to buf. * Use value of table_name FDW option (if any) instead of relation's name. * Similarly, schema_name FDW option overrides schema name. */ static void deparseRelation(StringInfo buf, Relation rel, QueryEngineType query_engine_type) { ForeignTable *table; const char *nspname = NULL; const char *relname = NULL; ListCell *lc; /* obtain additional catalog information. */ table = GetForeignTable(RelationGetRelid(rel)); /* * Use value of FDW options if any, instead of the name of object itself. */ foreach(lc, table->options) { DefElem *def = (DefElem *) lfirst(lc); if (strcmp(def->defname, "table") == 0) relname = defGetString(def); } /* * Note: we could skip printing the schema name if it's pg_catalog, but * that doesn't seem worth the trouble. */ if (nspname == NULL) nspname = get_namespace_name(RelationGetNamespace(rel)); if (relname == NULL) relname = RelationGetRelationName(rel); appendStringInfo(buf, "%s", td_quote_identifier(relname, query_engine_type)); } /* * Append a SQL string literal representing "val" to buf. */ void deparseStringLiteral(StringInfo buf, const char *val) { deparseStringLiteralForQueryEngineType(buf, val, QUERY_ENGINE_PRESTO); } static void deparseStringLiteralForQueryEngineType(StringInfo buf, const char *val, QueryEngineType query_engine_type) { const char *valptr; /* * Rather than making assumptions about the remote server's value of * standard_conforming_strings, always use E'foo' syntax if there are any * backslashes. This will fail on remote servers before 8.1, but those * are long out of support. */ if (strchr(val, '\\') != NULL) appendStringInfoChar(buf, ESCAPE_STRING_SYNTAX); appendStringInfoChar(buf, '\''); for (valptr = val; *valptr; valptr++) { char ch = *valptr; if (SQL_STR_DOUBLE(ch, true)) appendStringInfoChar(buf, ch); /* Workaround for https://issues.apache.org/jira/browse/HIVE-2409 */ if (query_engine_type == QUERY_ENGINE_HIVE && ch == ';') appendStringInfoChar(buf, '\\'); appendStringInfoChar(buf, ch); } appendStringInfoChar(buf, '\''); } /* * Deparse given expression into context->buf. * * This function must support all the same node types that foreign_expr_walker * accepts. * * Note: unlike ruleutils.c, we just use a simple hard-wired parenthesization * scheme: anything more complex than a Var, Const, function call or cast * should be self-parenthesized. */ static void deparseExpr(Expr *node, deparse_expr_cxt *context) { if (node == NULL) return; switch (nodeTag(node)) { case T_Var: deparseVar((Var *) node, context); break; case T_Const: deparseConst((Const *) node, context); break; case T_Param: deparseParam((Param *) node, context); break; case T_ArrayRef: deparseArrayRef((ArrayRef *) node, context); break; case T_FuncExpr: deparseFuncExpr((FuncExpr *) node, context); break; case T_OpExpr: deparseOpExpr((OpExpr *) node, context); break; case T_DistinctExpr: deparseDistinctExpr((DistinctExpr *) node, context); break; case T_ScalarArrayOpExpr: deparseScalarArrayOpExpr((ScalarArrayOpExpr *) node, context); break; case T_RelabelType: deparseRelabelType((RelabelType *) node, context); break; case T_BoolExpr: deparseBoolExpr((BoolExpr *) node, context); break; case T_NullTest: deparseNullTest((NullTest *) node, context); break; case T_ArrayExpr: deparseArrayExpr((ArrayExpr *) node, context); break; default: elog(ERROR, "unsupported expression type for deparse: %d", (int) nodeTag(node)); break; } } /* * Deparse given Var node into context->buf. * * If the Var belongs to the foreign relation, just print its remote name. * Otherwise, it's effectively a Param (and will in fact be a Param at * run time). Handle it the same way we handle plain Params --- see * deparseParam for comments. */ static void deparseVar(Var *node, deparse_expr_cxt *context) { StringInfo buf = context->buf; if (node->varno == context->foreignrel->relid && node->varlevelsup == 0) { /* Var belongs to foreign table */ deparseColumnRef(buf, node->varno, node->varattno, context->root, context->query_engine_type); } else { /* Treat like a Param */ if (context->params_list) { int pindex = 0; ListCell *lc; /* find its index in params_list */ foreach(lc, *context->params_list) { pindex++; if (equal(node, (Node *) lfirst(lc))) break; } if (lc == NULL) { /* not in list, so add it */ pindex++; *context->params_list = lappend(*context->params_list, node); } printRemoteParam(pindex, node->vartype, node->vartypmod, context); } else { printRemotePlaceholder(node->vartype, node->vartypmod, context); } } } /* * Deparse given constant value into context->buf. * * This function has to be kept in sync with ruleutils.c's get_const_expr. */ static void deparseConst(Const *node, deparse_expr_cxt *context) { StringInfo buf = context->buf; Oid typoutput; bool typIsVarlena; char *extval; #if 0 bool isfloat = false; bool needlabel; #endif if (node->constisnull) { appendStringInfoString(buf, "NULL"); #if 0 appendStringInfo(buf, "::%s", format_type_with_typemod(node->consttype, node->consttypmod)); #endif return; } getTypeOutputInfo(node->consttype, &typoutput, &typIsVarlena); extval = OidOutputFunctionCall(typoutput, node->constvalue); switch (node->consttype) { case INT2OID: case INT4OID: case INT8OID: case OIDOID: case FLOAT4OID: case FLOAT8OID: case NUMERICOID: { /* * No need to quote unless it's a special value such as 'NaN'. * See comments in get_const_expr(). */ if (strspn(extval, "0123456789+-eE.") == strlen(extval)) { if (extval[0] == '+' || extval[0] == '-') appendStringInfo(buf, "(%s)", extval); else appendStringInfoString(buf, extval); #if 0 if (strcspn(extval, "eE.") != strlen(extval)) isfloat = true; /* it looks like a float */ #endif } else appendStringInfo(buf, "'%s'", extval); } break; case BITOID: case VARBITOID: appendStringInfo(buf, "B'%s'", extval); break; case BOOLOID: if (strcmp(extval, "t") == 0) appendStringInfoString(buf, "true"); else appendStringInfoString(buf, "false"); break; default: deparseStringLiteralForQueryEngineType(buf, extval, context->query_engine_type); break; } #if 0 /* * Append ::typename unless the constant will be implicitly typed as the * right type when it is read in. * * XXX this code has to be kept in sync with the behavior of the parser, * especially make_const. */ switch (node->consttype) { case BOOLOID: case INT4OID: case UNKNOWNOID: needlabel = false; break; case NUMERICOID: needlabel = !isfloat || (node->consttypmod >= 0); break; default: needlabel = true; break; } if (needlabel) appendStringInfo(buf, "::%s", format_type_with_typemod(node->consttype, node->consttypmod)); #endif } /* * Deparse given Param node. * * If we're generating the query "for real", add the Param to * context->params_list if it's not already present, and then use its index * in that list as the remote parameter number. During EXPLAIN, there's * no need to identify a parameter number. */ static void deparseParam(Param *node, deparse_expr_cxt *context) { if (context->params_list) { int pindex = 0; ListCell *lc; /* find its index in params_list */ foreach(lc, *context->params_list) { pindex++; if (equal(node, (Node *) lfirst(lc))) break; } if (lc == NULL) { /* not in list, so add it */ pindex++; *context->params_list = lappend(*context->params_list, node); } printRemoteParam(pindex, node->paramtype, node->paramtypmod, context); } else { printRemotePlaceholder(node->paramtype, node->paramtypmod, context); } } /* * Deparse an array subscript expression. */ static void deparseArrayRef(ArrayRef *node, deparse_expr_cxt *context) { StringInfo buf = context->buf; ListCell *lowlist_item; ListCell *uplist_item; /* Always parenthesize the expression. */ appendStringInfoChar(buf, '('); /* * Deparse referenced array expression first. If that expression includes * a cast, we have to parenthesize to prevent the array subscript from * being taken as typename decoration. We can avoid that in the typical * case of subscripting a Var, but otherwise do it. */ if (IsA(node->refexpr, Var)) deparseExpr(node->refexpr, context); else { appendStringInfoChar(buf, '('); deparseExpr(node->refexpr, context); appendStringInfoChar(buf, ')'); } /* Deparse subscript expressions. */ lowlist_item = list_head(node->reflowerindexpr); /* could be NULL */ foreach(uplist_item, node->refupperindexpr) { appendStringInfoChar(buf, '['); if (lowlist_item) { deparseExpr(lfirst(lowlist_item), context); appendStringInfoChar(buf, ':'); lowlist_item = lnext(lowlist_item); } deparseExpr(lfirst(uplist_item), context); appendStringInfoChar(buf, ']'); } appendStringInfoChar(buf, ')'); } /* * Deparse a function call. */ static void deparseFuncExpr(FuncExpr *node, deparse_expr_cxt *context) { StringInfo buf = context->buf; HeapTuple proctup; Form_pg_proc procform; const char *proname; bool use_variadic; bool first; ListCell *arg; /* * If the function call came from an implicit coercion, then just show the * first argument. */ if (node->funcformat == COERCE_IMPLICIT_CAST) { deparseExpr((Expr *) linitial(node->args), context); return; } /* * If the function call came from a cast, then show the first argument * plus an explicit cast operation. */ if (node->funcformat == COERCE_EXPLICIT_CAST) { #if 0 Oid rettype = node->funcresulttype; #endif int32 coercedTypmod; /* Get the typmod if this is a length-coercion function */ (void) exprIsLengthCoercion((Node *) node, &coercedTypmod); deparseExpr((Expr *) linitial(node->args), context); #if 0 appendStringInfo(buf, "::%s", format_type_with_typemod(rettype, coercedTypmod)); #endif return; } /* * Normal function: display as proname(args). */ proctup = SearchSysCache1(PROCOID, ObjectIdGetDatum(node->funcid)); if (!HeapTupleIsValid(proctup)) elog(ERROR, "cache lookup failed for function %u", node->funcid); procform = (Form_pg_proc) GETSTRUCT(proctup); /* Check if need to print VARIADIC (cf. ruleutils.c) */ use_variadic = node->funcvariadic; /* Deparse the function name ... */ proname = NameStr(procform->proname); appendStringInfo(buf, "%s(", proname); /* ... and all the arguments */ first = true; foreach(arg, node->args) { if (!first) appendStringInfoString(buf, ", "); if (use_variadic && lnext(arg) == NULL) appendStringInfoString(buf, "VARIADIC "); deparseExpr((Expr *) lfirst(arg), context); first = false; } appendStringInfoChar(buf, ')'); ReleaseSysCache(proctup); } /* * Deparse given operator expression. To avoid problems around * priority of operations, we always parenthesize the arguments. */ static void deparseOpExpr(OpExpr *node, deparse_expr_cxt *context) { StringInfo buf = context->buf; HeapTuple tuple; Form_pg_operator form; char oprkind; ListCell *arg; /* Retrieve information about the operator from system catalog. */ tuple = SearchSysCache1(OPEROID, ObjectIdGetDatum(node->opno)); if (!HeapTupleIsValid(tuple)) elog(ERROR, "cache lookup failed for operator %u", node->opno); form = (Form_pg_operator) GETSTRUCT(tuple); oprkind = form->oprkind; /* Sanity check. */ Assert((oprkind == 'r' && list_length(node->args) == 1) || (oprkind == 'l' && list_length(node->args) == 1) || (oprkind == 'b' && list_length(node->args) == 2)); /* Hive doesn't support '||' operator */ if (context->query_engine_type == QUERY_ENGINE_HIVE) { char *opname = NameStr(form->oprname); if (oprkind == 'b' && strcmp(opname, "||") == 0) { appendStringInfoString(buf, "CONCAT("); /* Deparse left operand */ arg = list_head(node->args); deparseExpr(lfirst(arg), context); appendStringInfoString(buf, ", "); /* Deparse right operand */ arg = list_tail(node->args); deparseExpr(lfirst(arg), context); appendStringInfoChar(buf, ')'); ReleaseSysCache(tuple); return; } } /* Always parenthesize the expression. */ appendStringInfoChar(buf, '('); /* Deparse left operand. */ if (oprkind == 'r' || oprkind == 'b') { arg = list_head(node->args); deparseExpr(lfirst(arg), context); appendStringInfoChar(buf, ' '); } /* Deparse operator name. */ deparseOperatorName(buf, form); /* Deparse right operand. */ if (oprkind == 'l' || oprkind == 'b') { arg = list_tail(node->args); appendStringInfoChar(buf, ' '); deparseExpr(lfirst(arg), context); } appendStringInfoChar(buf, ')'); ReleaseSysCache(tuple); } /* * Print the name of an operator. */ static void deparseOperatorName(StringInfo buf, Form_pg_operator opform) { char *opname; /* opname is not a SQL identifier, so we should not quote it. */ opname = NameStr(opform->oprname); if (strcmp(opname, "~~") == 0) { appendStringInfoString(buf, "LIKE"); } else if (strcmp(opname, "!~~") == 0) { appendStringInfoString(buf, "NOT LIKE"); } else { appendStringInfoString(buf, opname); } } /* * Deparse IS DISTINCT FROM. */ static void deparseDistinctExpr(DistinctExpr *node, deparse_expr_cxt *context) { StringInfo buf = context->buf; Assert(list_length(node->args) == 2); appendStringInfoChar(buf, '('); deparseExpr(linitial(node->args), context); appendStringInfoString(buf, " IS DISTINCT FROM "); deparseExpr(lsecond(node->args), context); appendStringInfoChar(buf, ')'); } /* * Deparse given ScalarArrayOpExpr expression. To avoid problems * around priority of operations, we always parenthesize the arguments. */ static void deparseScalarArrayOpExpr(ScalarArrayOpExpr *node, deparse_expr_cxt *context) { StringInfo buf = context->buf; HeapTuple tuple; Form_pg_operator form; Expr *arg1; Expr *arg2; char *opname; /* Retrieve information about the operator from system catalog. */ tuple = SearchSysCache1(OPEROID, ObjectIdGetDatum(node->opno)); if (!HeapTupleIsValid(tuple)) elog(ERROR, "cache lookup failed for operator %u", node->opno); form = (Form_pg_operator) GETSTRUCT(tuple); /* Sanity check. */ Assert(list_length(node->args) == 2); opname = NameStr(form->oprname); /* Get arguments */ arg1 = linitial(node->args); arg2 = lsecond(node->args); switch (nodeTag((Node*)arg2)) { case T_Const: { Const *c; /* Deparse left operand. */ deparseExpr(arg1, context); /* Deparse right operand. */ c = (Const*)arg2; if (strcmp(opname, "<>") == 0) appendStringInfo(buf, " NOT"); appendStringInfo(buf, " IN ("); if (!c->constisnull) { Oid typoutput; bool typIsVarlena; char *extval; getTypeOutputInfo(c->consttype, &typoutput, &typIsVarlena); extval = OidOutputFunctionCall(typoutput, c->constvalue); switch (c->consttype) { case INT4ARRAYOID: case OIDARRAYOID: td_deparse_string(buf, extval, false, context->query_engine_type); break; default: td_deparse_string(buf, extval, true, context->query_engine_type); break; } } else { appendStringInfoString(buf, "NULL"); } appendStringInfoChar(buf, ')'); } break; default: if (context->query_engine_type == QUERY_ENGINE_HIVE) { appendStringInfoString(buf, "array_contains("); } else { appendStringInfoString(buf, "contains("); } /* Deparse right operand. */ deparseExpr(arg2, context); appendStringInfoString(buf, ", "); /* Deparse left operand. */ deparseExpr(arg1, context); appendStringInfoChar(buf, ')'); break; } ReleaseSysCache(tuple); } /* * Deparse a RelabelType (binary-compatible cast) node. */ static void deparseRelabelType(RelabelType *node, deparse_expr_cxt *context) { deparseExpr(node->arg, context); #if 0 if (node->relabelformat != COERCE_IMPLICIT_CAST) appendStringInfo(context->buf, "::%s", format_type_with_typemod(node->resulttype, node->resulttypmod)); #endif } /* * Deparse a BoolExpr node. */ static void deparseBoolExpr(BoolExpr *node, deparse_expr_cxt *context) { StringInfo buf = context->buf; const char *op = NULL; /* keep compiler quiet */ bool first; ListCell *lc; switch (node->boolop) { case AND_EXPR: op = "AND"; break; case OR_EXPR: op = "OR"; break; case NOT_EXPR: appendStringInfoString(buf, "(NOT "); deparseExpr(linitial(node->args), context); appendStringInfoChar(buf, ')'); return; } appendStringInfoChar(buf, '('); first = true; foreach(lc, node->args) { if (!first) appendStringInfo(buf, " %s ", op); deparseExpr((Expr *) lfirst(lc), context); first = false; } appendStringInfoChar(buf, ')'); } /* * Deparse IS [NOT] NULL expression. */ static void deparseNullTest(NullTest *node, deparse_expr_cxt *context) { StringInfo buf = context->buf; appendStringInfoChar(buf, '('); deparseExpr(node->arg, context); /* * For scalar inputs, we prefer to print as IS [NOT] NULL, which is * shorter and traditional. If it's a rowtype input but we're applying a * scalar test, must print IS [NOT] DISTINCT FROM NULL to be semantically * correct. */ if (node->argisrow || !type_is_rowtype(exprType((Node *) node->arg))) { if (node->nulltesttype == IS_NULL) appendStringInfoString(buf, " IS NULL)"); else appendStringInfoString(buf, " IS NOT NULL)"); } else { if (node->nulltesttype == IS_NULL) appendStringInfoString(buf, " IS NOT DISTINCT FROM NULL)"); else appendStringInfoString(buf, " IS DISTINCT FROM NULL)"); } } /* * Deparse ARRAY[...] construct. */ static void deparseArrayExpr(ArrayExpr *node, deparse_expr_cxt *context) { StringInfo buf = context->buf; bool first = true; ListCell *lc; appendStringInfoString(buf, "ARRAY["); foreach(lc, node->elements) { if (!first) appendStringInfoString(buf, ", "); deparseExpr(lfirst(lc), context); first = false; } appendStringInfoChar(buf, ']'); #if 0 /* If the array is empty, we need an explicit cast to the array type. */ if (node->elements == NIL) appendStringInfo(buf, "::%s", format_type_with_typemod(node->array_typeid, -1)); #endif } /* * Print the representation of a parameter to be sent to the remote side. * * Note: we always label the Param's type explicitly rather than relying on * transmitting a numeric type OID in PQexecParams(). This allows us to * avoid assuming that types have the same OIDs on the remote side as they * do locally --- they need only have the same names. */ static void printRemoteParam(int paramindex, Oid paramtype, int32 paramtypmod, deparse_expr_cxt *context) { StringInfo buf = context->buf; #if 0 char *ptypename = format_type_with_typemod(paramtype, paramtypmod); appendStringInfo(buf, "$%d::%s", paramindex, ptypename); #endif appendStringInfo(buf, "$%d", paramindex); } /* * Print the representation of a placeholder for a parameter that will be * sent to the remote side at execution time. * * This is used when we're just trying to EXPLAIN the remote query. * We don't have the actual value of the runtime parameter yet, and we don't * want the remote planner to generate a plan that depends on such a value * anyway. Thus, we can't do something simple like "$1::paramtype". * Instead, we emit "((SELECT null::paramtype)::paramtype)". * In all extant versions of Postgres, the planner will see that as an unknown * constant value, which is what we want. This might need adjustment if we * ever make the planner flatten scalar subqueries. Note: the reason for the * apparently useless outer cast is to ensure that the representation as a * whole will be parsed as an a_expr and not a select_with_parens; the latter * would do the wrong thing in the context "x = ANY(...)". */ static void printRemotePlaceholder(Oid paramtype, int32 paramtypmod, deparse_expr_cxt *context) { StringInfo buf = context->buf; char *ptypename = format_type_with_typemod(paramtype, paramtypmod); appendStringInfo(buf, "((SELECT null::%s)::%s)", ptypename, ptypename); } /* * This table includes function names that work on PostgreSQL and Hive/Presto * on Treasure Data. */ char allowed_functions[][16] = { /* Aggregation functions */ "avg", "count", "covar_pop", "covar_samp", "max", "min", "stddev_pop", "stddev_samp", "sum", "var_pop", "var_samp", "variance", /* Mathmatical functions */ "abs", "acos", "asin", "atan", "cbrt", "ceil", "ceiling", "cos", "degrees", "exp", "floor", "ln", "log", "pi", "power", "radians", "round", "sign", "sin", "sqrt", "tan", /* String functions */ "chr", "concat", "length", "lower", "lpad", "ltrim", "reverse", "rpad", "rtrim", "substr", "substring", "trim", "upper", "" }; static bool isAllowedFunction(FuncExpr *node) { int i; HeapTuple proctup; Form_pg_proc procform; char *allowed_funcname; const char *proname; bool is_found = false; if (node->funcformat == COERCE_IMPLICIT_CAST) { elog(DEBUG1, "This function is COERCE_IMPLICIT_CAST"); return true; } else if (node->funcformat == COERCE_EXPLICIT_CAST) { elog(DEBUG1, "This function is COERCE_EXPLICIT_CAST"); return true; } /* * Normal function: display as proname(args). */ proctup = SearchSysCache1(PROCOID, ObjectIdGetDatum(node->funcid)); if (!HeapTupleIsValid(proctup)) elog(ERROR, "cache lookup failed for function %u", node->funcid); procform = (Form_pg_proc) GETSTRUCT(proctup); /* Print schema name only if it's not pg_catalog */ if (procform->pronamespace != PG_CATALOG_NAMESPACE) { elog(DEBUG1, "This function isn't in pg_catalog"); } else { proname = NameStr(procform->proname); for (i = 0; ; i++) { allowed_funcname = allowed_functions[i]; if (strlen(allowed_funcname) == 0) { break; } if (strcmp(proname, allowed_funcname) == 0) { is_found = true; break; } } if (is_found) elog(DEBUG1, "This function works on Treasure Data: %s", proname); else elog(DEBUG1, "This function doesn't work on Treasure Data: %s", proname); } ReleaseSysCache(proctup); return is_found; } static char * td_quote_identifier(const char *s, QueryEngineType query_engine_type) { if (query_engine_type == QUERY_ENGINE_PRESTO) return (char *) quote_identifier(s); { char *result = palloc(strlen(s) * 2 + 3); char *r = result; *r++ = '`'; while (*s) { if (*s == '`') *r++ = *s; *r++ = *s; s++; } *r++ = '`'; *r++ = '\0'; return result; } } static void td_deparse_string(StringInfo buf, const char *val, bool is_str, QueryEngineType query_engine_type) { const char *valptr; int i = -1; bool in_quotes = false; bool in_escape_seq = false; if (is_str) appendStringInfoChar(buf, '\''); for (valptr = val; *valptr; valptr++) { char ch = *valptr; i++; if (in_escape_seq) { /* If the previous character is an escape character, just output this character as is */ in_escape_seq = false; appendStringInfoChar(buf, ch); continue; } switch (ch) { case '{': /* Remove head '{' because this syntax is not recognize by the remote server */ if (i != 0) appendStringInfoChar(buf, ch); break; case '}': /* Remove tail '}' because this syntax is not recognize by the remote server */ if (i != (strlen(val) - 1)) appendStringInfoChar(buf, ch); break; case '\"': /* Consider if current character is in a string */ in_quotes = !in_quotes; break; case ',': /* Handle it as a delimiter when not in a quoted string */ if (is_str && !in_quotes) { appendStringInfoChar(buf, '\''); appendStringInfoChar(buf, ch); appendStringInfoChar(buf, ' '); appendStringInfoChar(buf, '\''); continue; } else appendStringInfoChar(buf, ch); break; case ';': /* Workaround for https://issues.apache.org/jira/browse/HIVE-2409 */ if (query_engine_type == QUERY_ENGINE_HIVE) appendStringInfoChar(buf, '\\'); appendStringInfoChar(buf, ch); break; case '\\': in_escape_seq = true; break; case '\'': if (query_engine_type == QUERY_ENGINE_HIVE) appendStringInfoChar(buf, '\\'); else appendStringInfoChar(buf, '\''); appendStringInfoChar(buf, ch); break; default: appendStringInfoChar(buf, ch); break; } } if (is_str) appendStringInfoChar(buf, '\''); }