Add record_image_ops opclass for matview concurrent refresh.

REFRESH MATERIALIZED VIEW CONCURRENTLY was broken for any matview
containing a column of a type without a default btree operator
class.  It also did not produce results consistent with a non-
concurrent REFRESH or a normal view if any column was of a type
which allowed user-visible differences between values which
compared as equal according to the type's default btree opclass.
Concurrent matview refresh was modified to use the new operators
to solve these problems.

Documentation was added for record comparison, both for the
default btree operator class for record, and the newly added
operators.  Regression tests now check for proper behavior both
for a matview with a box column and a matview containing a citext
column.

Reviewed by Steve Singer, who suggested some of the doc language.

Author: kgrittn

contrib/citext/expected/citext.out

@@ -2276,3 +2276,44 @@ SELECT like_escape( name::text, ''::citext ) = like_escape( name::text, '' ) AS
  t
 (5 rows)
 
+-- Ensure correct behavior for citext with materialized views.
+CREATE TABLE citext_table (
+  id serial primary key,
+  name citext
+);
+INSERT INTO citext_table (name)
+  VALUES ('one'), ('two'), ('three'), (NULL), (NULL);
+CREATE MATERIALIZED VIEW citext_matview AS
+  SELECT * FROM citext_table;
+CREATE UNIQUE INDEX citext_matview_id
+  ON citext_matview (id);
+SELECT *
+  FROM citext_matview m
+  FULL JOIN citext_table t ON (t.id = m.id AND t *= m)
+  WHERE t.id IS NULL OR m.id IS NULL;
+ id | name | id | name 
+----+------+----+------
+(0 rows)
+
+UPDATE citext_table SET name = 'Two' WHERE name = 'TWO';
+SELECT *
+  FROM citext_matview m
+  FULL JOIN citext_table t ON (t.id = m.id AND t *= m)
+  WHERE t.id IS NULL OR m.id IS NULL;
+ id | name | id | name 
+----+------+----+------
+    |      |  2 | Two
+  2 | two  |    | 
+(2 rows)
+
+REFRESH MATERIALIZED VIEW CONCURRENTLY citext_matview;
+SELECT * FROM citext_matview ORDER BY id;
+ id | name  
+----+-------
+  1 | one
+  2 | Two
+  3 | three
+  4 | 
+  5 | 
+(5 rows)
+

contrib/citext/expected/citext_1.out

@@ -2276,3 +2276,44 @@ SELECT like_escape( name::text, ''::citext ) = like_escape( name::text, '' ) AS
  t
 (5 rows)
 
+-- Ensure correct behavior for citext with materialized views.
+CREATE TABLE citext_table (
+  id serial primary key,
+  name citext
+);
+INSERT INTO citext_table (name)
+  VALUES ('one'), ('two'), ('three'), (NULL), (NULL);
+CREATE MATERIALIZED VIEW citext_matview AS
+  SELECT * FROM citext_table;
+CREATE UNIQUE INDEX citext_matview_id
+  ON citext_matview (id);
+SELECT *
+  FROM citext_matview m
+  FULL JOIN citext_table t ON (t.id = m.id AND t *= m)
+  WHERE t.id IS NULL OR m.id IS NULL;
+ id | name | id | name 
+----+------+----+------
+(0 rows)
+
+UPDATE citext_table SET name = 'Two' WHERE name = 'TWO';
+SELECT *
+  FROM citext_matview m
+  FULL JOIN citext_table t ON (t.id = m.id AND t *= m)
+  WHERE t.id IS NULL OR m.id IS NULL;
+ id | name | id | name 
+----+------+----+------
+    |      |  2 | Two
+  2 | two  |    | 
+(2 rows)
+
+REFRESH MATERIALIZED VIEW CONCURRENTLY citext_matview;
+SELECT * FROM citext_matview ORDER BY id;
+ id | name  
+----+-------
+  1 | one
+  2 | Two
+  3 | three
+  4 | 
+  5 | 
+(5 rows)
+

contrib/citext/sql/citext.sql

@@ -711,3 +711,26 @@ SELECT COUNT(*) = 19::bigint AS t FROM try;
 
 SELECT like_escape( name, '' ) = like_escape( name::text, '' ) AS t FROM srt;
 SELECT like_escape( name::text, ''::citext ) = like_escape( name::text, '' ) AS t FROM srt;
+
+-- Ensure correct behavior for citext with materialized views.
+CREATE TABLE citext_table (
+  id serial primary key,
+  name citext
+);
+INSERT INTO citext_table (name)
+  VALUES ('one'), ('two'), ('three'), (NULL), (NULL);
+CREATE MATERIALIZED VIEW citext_matview AS
+  SELECT * FROM citext_table;
+CREATE UNIQUE INDEX citext_matview_id
+  ON citext_matview (id);
+SELECT *
+  FROM citext_matview m
+  FULL JOIN citext_table t ON (t.id = m.id AND t *= m)
+  WHERE t.id IS NULL OR m.id IS NULL;
+UPDATE citext_table SET name = 'Two' WHERE name = 'TWO';
+SELECT *
+  FROM citext_matview m
+  FULL JOIN citext_table t ON (t.id = m.id AND t *= m)
+  WHERE t.id IS NULL OR m.id IS NULL;
+REFRESH MATERIALIZED VIEW CONCURRENTLY citext_matview;
+SELECT * FROM citext_matview ORDER BY id;

doc/src/sgml/func.sgml

@@ -12739,7 +12739,7 @@ WHERE EXISTS (SELECT 1 FROM tab2 WHERE col2 = tab1.col2);
 
   <para>
    See <xref linkend="row-wise-comparison"> for details about the meaning
-   of a row-wise comparison.
+   of a row constructor comparison.
   </para>
   </sect2>
 
@@ -12795,12 +12795,12 @@ WHERE EXISTS (SELECT 1 FROM tab2 WHERE col2 = tab1.col2);
 
   <para>
    See <xref linkend="row-wise-comparison"> for details about the meaning
-   of a row-wise comparison.
+   of a row constructor comparison.
   </para>
   </sect2>
 
   <sect2>
-   <title>Row-wise Comparison</title>
+   <title>Single-row Comparison</title>
 
    <indexterm zone="functions-subquery">
     <primary>comparison</primary>
@@ -12823,7 +12823,7 @@ WHERE EXISTS (SELECT 1 FROM tab2 WHERE col2 = tab1.col2);
 
   <para>
    See <xref linkend="row-wise-comparison"> for details about the meaning
-   of a row-wise comparison.
+   of a row constructor comparison.
   </para>
   </sect2>
  </sect1>
@@ -12852,13 +12852,23 @@ WHERE EXISTS (SELECT 1 FROM tab2 WHERE col2 = tab1.col2);
    <primary>SOME</primary>
   </indexterm>
 
+  <indexterm>
+   <primary>composite type</primary>
+   <secondary>comparison</secondary>
+  </indexterm>
+
   <indexterm>
    <primary>row-wise comparison</primary>
   </indexterm>
 
   <indexterm>
    <primary>comparison</primary>
-   <secondary>row-wise</secondary>
+   <secondary>composite type</secondary>
+  </indexterm>
+
+  <indexterm>
+   <primary>comparison</primary>
+   <secondary>row constructor</secondary>
   </indexterm>
 
   <indexterm>
@@ -13023,7 +13033,7 @@ AND
   </sect2>
 
   <sect2 id="row-wise-comparison">
-   <title>Row-wise Comparison</title>
+   <title>Row Constructor Comparison</title>
 
 <synopsis>
 <replaceable>row_constructor</replaceable> <replaceable>operator</replaceable> <replaceable>row_constructor</replaceable>
@@ -13033,20 +13043,25 @@ AND
    Each side is a row constructor,
    as described in <xref linkend="sql-syntax-row-constructors">.
    The two row values must have the same number of fields.
-   Each side is evaluated and they are compared row-wise.  Row comparisons
-   are allowed when the <replaceable>operator</replaceable> is
+   Each side is evaluated and they are compared row-wise.  Row constructor
+   comparisons are allowed when the <replaceable>operator</replaceable> is
    <literal>=</>,
    <literal>&lt;&gt;</>,
    <literal>&lt;</>,
    <literal>&lt;=</>,
    <literal>&gt;</> or
-   <literal>&gt;=</>,
-   or has semantics similar to one of these.  (To be specific, an operator
-   can be a row comparison operator if it is a member of a B-tree operator
-   class, or is the negator of the <literal>=</> member of a B-tree operator
-   class.)
+   <literal>&gt;=</>.
+   Every row element must be of a type which has a default B-tree operator
+   class or the attempted comparison may generate an error.
   </para>
 
+  <note>
+   <para>
+    Errors related to the number or types of elements might not occur if
+    the comparison is resolved using earlier columns.
+   </para>
+  </note>
+
   <para>
    The <literal>=</> and <literal>&lt;&gt;</> cases work slightly differently
    from the others.  Two rows are considered
@@ -13104,20 +13119,64 @@ AND
    be either true or false, never null.
   </para>
 
-  <note>
-   <para>
-    The SQL specification requires row-wise comparison to return NULL if the
-    result depends on comparing two NULL values or a NULL and a non-NULL.
-    <productname>PostgreSQL</productname> does this only when comparing the
-    results of two row constructors or comparing a row constructor to the
-    output of a subquery (as in <xref linkend="functions-subquery">).
-    In other contexts where two composite-type values are compared, two
-    NULL field values are considered equal, and a NULL is considered larger
-    than a non-NULL.  This is necessary in order to have consistent sorting
-    and indexing behavior for composite types.
-   </para>
-  </note>
+  </sect2>
 
+  <sect2 id="composite-type-comparison">
+   <title>Composite Type Comparison</title>
+
+<synopsis>
+<replaceable>record</replaceable> <replaceable>operator</replaceable> <replaceable>record</replaceable>
+</synopsis>
+
+  <para>
+   The SQL specification requires row-wise comparison to return NULL if the
+   result depends on comparing two NULL values or a NULL and a non-NULL.
+   <productname>PostgreSQL</productname> does this only when comparing the
+   results of two row constructors (as in
+   <xref linkend="row-wise-comparison">) or comparing a row constructor
+   to the output of a subquery (as in <xref linkend="functions-subquery">).
+   In other contexts where two composite-type values are compared, two
+   NULL field values are considered equal, and a NULL is considered larger
+   than a non-NULL.  This is necessary in order to have consistent sorting
+   and indexing behavior for composite types.
+  </para>
+
+  <para>
+   Each side is evaluated and they are compared row-wise.  Composite type
+   comparisons are allowed when the <replaceable>operator</replaceable> is
+   <literal>=</>,
+   <literal>&lt;&gt;</>,
+   <literal>&lt;</>,
+   <literal>&lt;=</>,
+   <literal>&gt;</> or
+   <literal>&gt;=</>,
+   or has semantics similar to one of these.  (To be specific, an operator
+   can be a row comparison operator if it is a member of a B-tree operator
+   class, or is the negator of the <literal>=</> member of a B-tree operator
+   class.)  The default behavior of the above operators is the same as for
+   <literal>IS [ NOT ] DISTINCT FROM</literal> for row constructors (see
+   <xref linkend="row-wise-comparison">).
+  </para>
+
+  <para>
+   To support matching of rows which include elements without a default
+   B-tree operator class, the following operators are defined for composite
+   type comparison:
+   <literal>*=</>,
+   <literal>*&lt;&gt;</>,
+   <literal>*&lt;</>,
+   <literal>*&lt;=</>,
+   <literal>*&gt;</>, and
+   <literal>*&gt;=</>.
+   These operators compare the internal binary representation of the two
+   rows.  Two rows might have a different binary representation even
+   though comparisons of the two rows with the equality operator is true. 
+   The ordering of rows under these comparision operators is deterministic
+   but not otherwise meaningful.  These operators are used internally for
+   materialized views and might be useful for other specialized purposes
+   such as replication but are not intended to be generally useful for
+   writing queries.
+  </para>
   </sect2>
  </sect1>
 

src/backend/commands/matview.c

@@ -562,7 +562,7 @@ refresh_by_match_merge(Oid matviewOid, Oid tempOid)
 					 "SELECT newdata FROM %s newdata "
 					 "WHERE newdata IS NOT NULL AND EXISTS "
 					 "(SELECT * FROM %s newdata2 WHERE newdata2 IS NOT NULL "
-					 "AND newdata2 OPERATOR(pg_catalog.=) newdata "
+					 "AND newdata2 OPERATOR(pg_catalog.*=) newdata "
 					 "AND newdata2.ctid OPERATOR(pg_catalog.<>) "
 					 "newdata.ctid) LIMIT 1",
 					 tempname, tempname);
@@ -645,9 +645,6 @@ refresh_by_match_merge(Oid matviewOid, Oid tempOid)
 				/*
 				 * Only include the column once regardless of how many times
 				 * it shows up in how many indexes.
-				 *
-				 * This is also useful later to omit columns which can not
-				 * have changed from the SET clause of the UPDATE statement.
 				 */
 				if (usedForQual[attnum - 1])
 					continue;
@@ -682,8 +679,9 @@ refresh_by_match_merge(Oid matviewOid, Oid tempOid)
 				 errhint("Create a UNIQUE index with no WHERE clause on one or more columns of the materialized view.")));
 
 	appendStringInfoString(&querybuf,
-						   " AND newdata = mv) WHERE newdata IS NULL OR mv IS NULL"
-						   " ORDER BY tid");
+						   " AND newdata OPERATOR(pg_catalog.*=) mv) "
+						   "WHERE newdata IS NULL OR mv IS NULL "
+						   "ORDER BY tid");
 
 	/* Create the temporary "diff" table. */
 	if (SPI_exec(querybuf.data, 0) != SPI_OK_UTILITY)