MERGE SQL Command following SQL:2016

MERGE performs actions that modify rows in the target table
using a source table or query. MERGE provides a single SQL
statement that can conditionally INSERT/UPDATE/DELETE rows
a task that would other require multiple PL statements.
e.g.

MERGE INTO target AS t
USING source AS s
ON t.tid = s.sid
WHEN MATCHED AND t.balance > s.delta THEN
  UPDATE SET balance = t.balance - s.delta
WHEN MATCHED THEN
  DELETE
WHEN NOT MATCHED AND s.delta > 0 THEN
  INSERT VALUES (s.sid, s.delta)
WHEN NOT MATCHED THEN
  DO NOTHING;

MERGE works with regular and partitioned tables, including
column and row security enforcement, as well as support for
row, statement and transition triggers.

MERGE is optimized for OLTP and is parameterizable, though
also useful for large scale ETL/ELT. MERGE is not intended
to be used in preference to existing single SQL commands
for INSERT, UPDATE or DELETE since there is some overhead.
MERGE can be used statically from PL/pgSQL.

MERGE does not yet support inheritance, write rules,
RETURNING clauses, updatable views or foreign tables.
MERGE follows SQL Standard per the most recent SQL:2016.

Includes full tests and documentation, including full
isolation tests to demonstrate the concurrent behavior.

This version written from scratch in 2017 by Simon Riggs,
using docs and tests originally written in 2009. Later work
from Pavan Deolasee has been both complex and deep, leaving
the lead author credit now in his hands.
Extensive discussion of concurrency from Peter Geoghegan,
with thanks for the time and effort contributed.

Various issues reported via sqlsmith by Andreas Seltenreich

Authors: Pavan Deolasee, Simon Riggs
Reviewer: Peter Geoghegan, Amit Langote, Tomas Vondra, Simon Riggs

Discussion:
https://postgr.es/m/CANP8+jKitBSrB7oTgT9CY2i1ObfOt36z0XMraQc+Xrz8QB0nXA@mail.gmail.com
https://postgr.es/m/CAH2-WzkJdBuxj9PO=2QaO9-3h3xGbQPZ34kJH=HukRekwM-GZg@mail.gmail.com

1 files changed, 40 insertions, 5 deletions

diff --git a/src/backend/parser/parse_clause.c b/src/backend/parser/parse_clause.c
index 3a02307bd9..3cb761b4ed 100644
--- a/src/backend/parser/parse_clause.c
+++ b/src/backend/parser/parse_clause.c
@@ -76,9 +76,6 @@ static RangeTblEntry *transformRangeTableFunc(ParseState *pstate,
 						RangeTableFunc *t);
 static TableSampleClause *transformRangeTableSample(ParseState *pstate,
 						  RangeTableSample *rts);
-static Node *transformFromClauseItem(ParseState *pstate, Node *n,
-						RangeTblEntry **top_rte, int *top_rti,
-						List **namespace);
 static Node *buildMergedJoinVar(ParseState *pstate, JoinType jointype,
 				   Var *l_colvar, Var *r_colvar);
 static ParseNamespaceItem *makeNamespaceItem(RangeTblEntry *rte,
@@ -139,6 +136,7 @@ transformFromClause(ParseState *pstate, List *frmList)
 		n = transformFromClauseItem(pstate, n,
 									&rte,
 									&rtindex,
+									NULL, NULL,
 									&namespace);
 
 		checkNameSpaceConflicts(pstate, pstate->p_namespace, namespace);
@@ -1096,13 +1094,20 @@ getRTEForSpecialRelationTypes(ParseState *pstate, RangeVar *rv)
  *
  * *top_rti: receives the rangetable index of top_rte.  (Ditto.)
  *
+ * *right_rte: receives the RTE corresponding to the right side of the
+ * jointree. Only MERGE really needs to know about this and only MERGE passes a
+ * non-NULL pointer.
+ *
+ * *right_rti: receives the rangetable index of the right_rte.
+ *
  * *namespace: receives a List of ParseNamespaceItems for the RTEs exposed
  * as table/column names by this item.  (The lateral_only flags in these items
  * are indeterminate and should be explicitly set by the caller before use.)
  */
-static Node *
+Node *
 transformFromClauseItem(ParseState *pstate, Node *n,
 						RangeTblEntry **top_rte, int *top_rti,
+						RangeTblEntry **right_rte, int *right_rti,
 						List **namespace)
 {
 	if (IsA(n, RangeVar))
@@ -1194,7 +1199,7 @@ transformFromClauseItem(ParseState *pstate, Node *n,
 
 		/* Recursively transform the contained relation */
 		rel = transformFromClauseItem(pstate, rts->relation,
-									  top_rte, top_rti, namespace);
+									  top_rte, top_rti, NULL, NULL, namespace);
 		/* Currently, grammar could only return a RangeVar as contained rel */
 		rtr = castNode(RangeTblRef, rel);
 		rte = rt_fetch(rtr->rtindex, pstate->p_rtable);
@@ -1222,6 +1227,7 @@ transformFromClauseItem(ParseState *pstate, Node *n,
 		List	   *l_namespace,
 				   *r_namespace,
 				   *my_namespace,
+				   *save_namespace,
 				   *l_colnames,
 				   *r_colnames,
 				   *res_colnames,
@@ -1240,6 +1246,7 @@ transformFromClauseItem(ParseState *pstate, Node *n,
 		j->larg = transformFromClauseItem(pstate, j->larg,
 										  &l_rte,
 										  &l_rtindex,
+										  NULL, NULL,
 										  &l_namespace);
 
 		/*
@@ -1263,12 +1270,34 @@ transformFromClauseItem(ParseState *pstate, Node *n,
 		sv_namespace_length = list_length(pstate->p_namespace);
 		pstate->p_namespace = list_concat(pstate->p_namespace, l_namespace);
 
+		/*
+		 * If we are running MERGE, don't make the other RTEs visible while
+		 * parsing the source relation. It mustn't see them.
+		 *
+		 * Currently, only MERGE passes non-NULL value for right_rte, so we
+		 * can safely deduce if we're running MERGE or not by just looking at
+		 * the right_rte. If that ever changes, we should look at other means
+		 * to find that.
+		 */
+		if (right_rte)
+		{
+			save_namespace = pstate->p_namespace;
+			pstate->p_namespace = NIL;
+		}
+
 		/* And now we can process the RHS */
 		j->rarg = transformFromClauseItem(pstate, j->rarg,
 										  &r_rte,
 										  &r_rtindex,
+										  NULL, NULL,
 										  &r_namespace);
 
+		/*
+		 * And now restore the namespace again so that join-quals can see it.
+		 */
+		if (right_rte)
+			pstate->p_namespace = save_namespace;
+
 		/* Remove the left-side RTEs from the namespace list again */
 		pstate->p_namespace = list_truncate(pstate->p_namespace,
 											sv_namespace_length);
@@ -1295,6 +1324,12 @@ transformFromClauseItem(ParseState *pstate, Node *n,
 		expandRTE(r_rte, r_rtindex, 0, -1, false,
 				  &r_colnames, &r_colvars);
 
+		if (right_rte)
+			*right_rte = r_rte;
+
+		if (right_rti)
+			*right_rti = r_rtindex;
+
 		/*
 		 * Natural join does not explicitly specify columns; must generate
 		 * columns to join. Need to run through the list of columns from each