Support assignment to subfields of composite columns in UPDATE and INSERT.

As a side effect, cause subscripts in INSERT targetlists to do something
more or less sensible; previously we evaluated such subscripts and then
effectively ignored them.  Another side effect is that UPDATE-ing an
element or slice of an array value that is NULL now produces a non-null
result, namely an array containing just the assigned-to positions.

5 files changed, 89 insertions, 35 deletions

diff --git a/src/include/nodes/execnodes.h b/src/include/nodes/execnodes.h
index 325bf876800..a32378c7f84 100644
--- a/src/include/nodes/execnodes.h
+++ b/src/include/nodes/execnodes.h
@@ -7,7 +7,7 @@
  * Portions Copyright (c) 1996-2003, PostgreSQL Global Development Group
  * Portions Copyright (c) 1994, Regents of the University of California
  *
- * $PostgreSQL: pgsql/src/include/nodes/execnodes.h,v 1.116 2004/05/10 22:44:49 tgl Exp $
+ * $PostgreSQL: pgsql/src/include/nodes/execnodes.h,v 1.117 2004/06/09 19:08:18 tgl Exp $
  *
  *-------------------------------------------------------------------------
  */
@@ -575,6 +575,18 @@ typedef struct FieldSelectState
 } FieldSelectState;
 
 /* ----------------
+ *		FieldStoreState node
+ * ----------------
+ */
+typedef struct FieldStoreState
+{
+	ExprState	xprstate;
+	ExprState  *arg;			/* input tuple value */
+	List	   *newvals;		/* new value(s) for field(s) */
+	TupleDesc	argdesc;		/* tupdesc for most recent input */
+} FieldStoreState;
+
+/* ----------------
  *		CaseExprState node
  * ----------------
  */
diff --git a/src/include/nodes/nodes.h b/src/include/nodes/nodes.h
index 6feedf6762c..b719747e762 100644
--- a/src/include/nodes/nodes.h
+++ b/src/include/nodes/nodes.h
@@ -7,7 +7,7 @@
  * Portions Copyright (c) 1996-2003, PostgreSQL Global Development Group
  * Portions Copyright (c) 1994, Regents of the University of California
  *
- * $PostgreSQL: pgsql/src/include/nodes/nodes.h,v 1.156 2004/05/26 13:57:02 momjian Exp $
+ * $PostgreSQL: pgsql/src/include/nodes/nodes.h,v 1.157 2004/06/09 19:08:18 tgl Exp $
  *
  *-------------------------------------------------------------------------
  */
@@ -110,6 +110,7 @@ typedef enum NodeTag
 	T_SubLink,
 	T_SubPlan,
 	T_FieldSelect,
+	T_FieldStore,
 	T_RelabelType,
 	T_CaseExpr,
 	T_CaseWhen,
@@ -143,6 +144,7 @@ typedef enum NodeTag
 	T_BoolExprState,
 	T_SubPlanState,
 	T_FieldSelectState,
+	T_FieldStoreState,
 	T_CaseExprState,
 	T_CaseWhenState,
 	T_ArrayExprState,
@@ -274,7 +276,7 @@ typedef enum NodeTag
 	T_A_Const,
 	T_FuncCall,
 	T_A_Indices,
-	T_ExprFieldSelect,
+	T_A_Indirection,
 	T_ResTarget,
 	T_TypeCast,
 	T_SortBy,
diff --git a/src/include/nodes/parsenodes.h b/src/include/nodes/parsenodes.h
index 8f6cc25e0aa..0da9b379077 100644
--- a/src/include/nodes/parsenodes.h
+++ b/src/include/nodes/parsenodes.h
@@ -7,7 +7,7 @@
  * Portions Copyright (c) 1996-2003, PostgreSQL Global Development Group
  * Portions Copyright (c) 1994, Regents of the University of California
  *
- * $PostgreSQL: pgsql/src/include/nodes/parsenodes.h,v 1.257 2004/06/02 21:01:09 momjian Exp $
+ * $PostgreSQL: pgsql/src/include/nodes/parsenodes.h,v 1.258 2004/06/09 19:08:18 tgl Exp $
  *
  *-------------------------------------------------------------------------
  */
@@ -166,26 +166,26 @@ typedef struct TypeName
  * ColumnRef - specifies a reference to a column, or possibly a whole tuple
  *
  *		The "fields" list must be nonempty; its last component may be "*"
- *		instead of a field name.  Subscripts are optional.
+ *		instead of a regular field name.
+ *
+ * Note: any array subscripting or selection of fields from composite columns
+ * is represented by an A_Indirection node above the ColumnRef.  However,
+ * for simplicity in the normal case, initial field selection from a table
+ * name is represented within ColumnRef and not by adding A_Indirection.
  */
 typedef struct ColumnRef
 {
 	NodeTag		type;
 	List	   *fields;			/* field names (list of Value strings) */
-	List	   *indirection;	/* subscripts (list of A_Indices) */
 } ColumnRef;
 
 /*
- * ParamRef - specifies a parameter reference
- *
- *		The parameter could be qualified with field names and/or subscripts
+ * ParamRef - specifies a $n parameter reference
  */
 typedef struct ParamRef
 {
 	NodeTag		type;
 	int			number;			/* the number of the parameter */
-	List	   *fields;			/* field names (list of Value strings) */
-	List	   *indirection;	/* subscripts (list of A_Indices) */
 } ParamRef;
 
 /*
@@ -267,40 +267,50 @@ typedef struct A_Indices
 } A_Indices;
 
 /*
- * ExprFieldSelect - select a field and/or array element from an expression
+ * A_Indirection - select a field and/or array element from an expression
  *
- *		This is used in the raw parsetree to represent selection from an
- *		arbitrary expression (not a column or param reference).  Either
- *		fields or indirection may be NIL if not used.
+ * The indirection list can contain both A_Indices nodes (representing
+ * subscripting) and string Value nodes (representing field selection
+ * --- the string value is the name of the field to select).  For example,
+ * a complex selection operation like
+ *				(foo).field1[42][7].field2
+ * would be represented with a single A_Indirection node having a 4-element
+ * indirection list.
+ *
+ * Note: as of Postgres 7.5, we don't support arrays of composite values,
+ * so cases in which a field select follows a subscript aren't actually
+ * semantically legal.  However the parser is prepared to handle such.
  */
-typedef struct ExprFieldSelect
+typedef struct A_Indirection
 {
 	NodeTag		type;
 	Node	   *arg;			/* the thing being selected from */
-	List	   *fields;			/* field names (list of Value strings) */
-	List	   *indirection;	/* subscripts (list of A_Indices) */
-} ExprFieldSelect;
+	List	   *indirection;	/* subscripts and/or field names */
+} A_Indirection;
 
 /*
  * ResTarget -
- *	  result target (used in target list of pre-transformed Parse trees)
+ *	  result target (used in target list of pre-transformed parse trees)
  *
- * In a SELECT or INSERT target list, 'name' is either NULL or
- * the column name assigned to the value.  (If there is an 'AS ColumnLabel'
- * clause, the grammar sets 'name' from it; otherwise 'name' is initially NULL
- * and is filled in during the parse analysis phase.)
- * The 'indirection' field is not used at all.
+ * In a SELECT or INSERT target list, 'name' is the column label from an
+ * 'AS ColumnLabel' clause, or NULL if there was none, and 'val' is the
+ * value expression itself.  The 'indirection' field is not used.
  *
- * In an UPDATE target list, 'name' is the name of the destination column,
+ * INSERT has a second ResTarget list which is the target-column-names list.
+ * Here, 'val' is not used, 'name' is the name of the destination column,
  * and 'indirection' stores any subscripts attached to the destination.
- * That is, our representation is UPDATE table SET name [indirection] = val.
+ *
+ * In an UPDATE target list, 'name' is the name of the destination column,
+ * 'indirection' stores any subscripts attached to the destination, and
+ * 'val' is the expression to assign.
+ *
+ * See A_Indirection for more info about what can appear in 'indirection'.
  */
 typedef struct ResTarget
 {
 	NodeTag		type;
 	char	   *name;			/* column name or NULL */
-	List	   *indirection;	/* subscripts for destination column, or
-								 * NIL */
+	List	   *indirection;	/* subscripts and field names, or NIL */
 	Node	   *val;			/* the value expression to compute or
 								 * assign */
 } ResTarget;
diff --git a/src/include/nodes/primnodes.h b/src/include/nodes/primnodes.h
index 5db00199cef..3ed4d74ee35 100644
--- a/src/include/nodes/primnodes.h
+++ b/src/include/nodes/primnodes.h
@@ -10,7 +10,7 @@
  * Portions Copyright (c) 1996-2003, PostgreSQL Global Development Group
  * Portions Copyright (c) 1994, Regents of the University of California
  *
- * $PostgreSQL: pgsql/src/include/nodes/primnodes.h,v 1.99 2004/05/30 23:40:39 neilc Exp $
+ * $PostgreSQL: pgsql/src/include/nodes/primnodes.h,v 1.100 2004/06/09 19:08:18 tgl Exp $
  *
  *-------------------------------------------------------------------------
  */
@@ -36,14 +36,15 @@
  * ordinal position (counting from 1).  However, in an INSERT or UPDATE
  * targetlist, resno represents the attribute number of the destination
  * column for the item; so there may be missing or out-of-order resnos.
- * In an UPDATE, it is even legal to have duplicated resnos; consider
+ * It is even legal to have duplicated resnos; consider
  *		UPDATE table SET arraycol[1] = ..., arraycol[2] = ..., ...
  * The two meanings come together in the executor, because the planner
  * transforms INSERT/UPDATE tlists into a normalized form with exactly
  * one entry for each column of the destination table.  Before that's
  * happened, however, it is risky to assume that resno == position.
  * Generally get_tle_by_resno() should be used rather than list_nth()
- * to fetch tlist entries by resno.
+ * to fetch tlist entries by resno, and only in SELECT should you assume
+ * that resno is a unique identifier.
  *
  * resname is required to represent the correct column name in non-resjunk
  * entries of top-level SELECT targetlists, since it will be used as the
@@ -541,6 +542,31 @@ typedef struct FieldSelect
 } FieldSelect;
 
 /* ----------------
+ * FieldStore
+ *
+ * FieldStore represents the operation of modifying one field in a tuple
+ * value, yielding a new tuple value (the input is not touched!).  Like
+ * the assign case of ArrayRef, this is used to implement UPDATE of a
+ * portion of a column.
+ *
+ * A single FieldStore can actually represent updates of several different
+ * fields.  The parser only generates FieldStores with single-element lists,
+ * but the planner will collapse multiple updates of the same base column
+ * into one FieldStore.
+ * ----------------
+ */
+
+typedef struct FieldStore
+{
+	Expr		xpr;
+	Expr	   *arg;			/* input tuple value */
+	List	   *newvals;		/* new value(s) for field(s) */
+	List	   *fieldnums;		/* integer list of field attnums */
+	Oid			resulttype;		/* type of result (same as type of arg) */
+	/* Like RowExpr, we deliberately omit a typmod here */
+} FieldStore;
+
+/* ----------------
  * RelabelType
  *
  * RelabelType represents a "dummy" type coercion between two binary-
@@ -607,6 +633,9 @@ typedef struct CaseWhen
  * Placeholder node for the test value to be processed by a CASE expression.
  * This is effectively like a Param, but can be implemented more simply
  * since we need only one replacement value at a time.
+ *
+ * We also use this in nested UPDATE expressions.
+ * See transformAssignmentIndirection().
  */
 typedef struct CaseTestExpr
 {
diff --git a/src/include/parser/parse_node.h b/src/include/parser/parse_node.h
index 8bb595c7098..de7b4766559 100644
--- a/src/include/parser/parse_node.h
+++ b/src/include/parser/parse_node.h
@@ -7,7 +7,7 @@
  * Portions Copyright (c) 1996-2003, PostgreSQL Global Development Group
  * Portions Copyright (c) 1994, Regents of the University of California
  *
- * $PostgreSQL: pgsql/src/include/parser/parse_node.h,v 1.38 2003/11/29 22:41:09 pgsql Exp $
+ * $PostgreSQL: pgsql/src/include/parser/parse_node.h,v 1.39 2004/06/09 19:08:19 tgl Exp $
  *
  *-------------------------------------------------------------------------
  */
@@ -67,12 +67,13 @@ typedef struct ParseState
 
 extern ParseState *make_parsestate(ParseState *parentParseState);
 extern Var *make_var(ParseState *pstate, RangeTblEntry *rte, int attrno);
+extern Oid transformArrayType(Oid arrayType);
 extern ArrayRef *transformArraySubscripts(ParseState *pstate,
 						 Node *arrayBase,
 						 Oid arrayType,
-						 int32 arrayTypMod,
+						 Oid elementType,
+						 int32 elementTypMod,
 						 List *indirection,
-						 bool forceSlice,
 						 Node *assignFrom);
 extern Const *make_const(Value *value);