v17-0002-alter-table-set-compression.patch

From ea7985a5e9589de86b7e680ac80fd4156aa16ee5 Mon Sep 17 00:00:00 2001
From: dilipkumar <dilipbalaut@gmail.com>
Date: Fri, 18 Dec 2020 11:31:22 +0530
Subject: [PATCH v17 2/6] alter table set compression

Add support for changing the compression method associated
with a column, forcing a table rewrite.

Dilip Kumar based on the patches from Ildus Kurbangaliev.
Design input from Robert Haas and Tomas Vondra.
Reviewed by Robert Haas, Tomas Vondra, Alexander Korotkov

Discussions:
https://www.postgresql.org/message-id/20171213151818.75a20259@postgrespro.ru
https://www.postgresql.org/message-id/CA%2BTgmoaKDW1Oi9V%3Djc9hOGyf77NbkNEABuqgHD1Cq%3D%3D1QsOcxg%40mail.gmail.com
https://www.postgresql.org/message-id/CA%2BTgmobSDVgUage9qQ5P_%3DF_9jaMkCgyKxUQGtFQU7oN4kX-AA%40mail.gmail.com
https://www.postgresql.org/message-id/20201005160355.byp74sh3ejsv7wrj%40development
https://www.postgresql.org/message-id/CAFiTN-tzTTT2oqWdRGLv1dvvS5MC1W%2BLE%2B3bqWPJUZj4GnHOJg%40mail.gmail.com
---
 doc/src/sgml/ref/alter_table.sgml         |  14 ++
 src/backend/commands/tablecmds.c          | 208 +++++++++++++++++-----
 src/backend/parser/gram.y                 |   9 +
 src/bin/psql/tab-complete.c               |   2 +-
 src/include/commands/event_trigger.h      |   1 +
 src/include/nodes/parsenodes.h            |   3 +-
 src/test/regress/expected/compression.out |  59 ++++++
 src/test/regress/sql/compression.sql      |  21 +++
 8 files changed, 272 insertions(+), 45 deletions(-)

diff --git a/doc/src/sgml/ref/alter_table.sgml b/doc/src/sgml/ref/alter_table.sgml
index c25ef5abd6..c4b53ec1c5 100644
--- a/doc/src/sgml/ref/alter_table.sgml
+++ b/doc/src/sgml/ref/alter_table.sgml
@@ -54,6 +54,7 @@ ALTER TABLE [ IF EXISTS ] <replaceable class="parameter">name</replaceable>
     ALTER [ COLUMN ] <replaceable class="parameter">column_name</replaceable> SET ( <replaceable class="parameter">attribute_option</replaceable> = <replaceable class="parameter">value</replaceable> [, ... ] )
     ALTER [ COLUMN ] <replaceable class="parameter">column_name</replaceable> RESET ( <replaceable class="parameter">attribute_option</replaceable> [, ... ] )
     ALTER [ COLUMN ] <replaceable class="parameter">column_name</replaceable> SET STORAGE { PLAIN | EXTERNAL | EXTENDED | MAIN }
+    ALTER [ COLUMN ] <replaceable class="parameter">column_name</replaceable> SET COMPRESSION <replaceable class="parameter">compression_method</replaceable>
     ADD <replaceable class="parameter">table_constraint</replaceable> [ NOT VALID ]
     ADD <replaceable class="parameter">table_constraint_using_index</replaceable>
     ALTER CONSTRAINT <replaceable class="parameter">constraint_name</replaceable> [ DEFERRABLE | NOT DEFERRABLE ] [ INITIALLY DEFERRED | INITIALLY IMMEDIATE ]
@@ -383,6 +384,19 @@ WITH ( MODULUS <replaceable class="parameter">numeric_literal</replaceable>, REM
     </listitem>
    </varlistentry>
 
+   <varlistentry>
+    <term>
+     <literal>SET COMPRESSION <replaceable class="parameter">compression_method</replaceable></literal>
+    </term>
+    <listitem>
+     <para>
+      This clause adds compression to a column. The Compression method can be
+      set to any available compression method.  The built-in methods are
+      <literal>pglz</literal> and <literal>lz4</literal>.
+     </para>
+    </listitem>
+   </varlistentry>
+
    <varlistentry>
     <term><literal>ADD <replaceable class="parameter">table_constraint</replaceable> [ NOT VALID ]</literal></term>
     <listitem>
diff --git a/src/backend/commands/tablecmds.c b/src/backend/commands/tablecmds.c
index fe33248589..8944ebc586 100644
--- a/src/backend/commands/tablecmds.c
+++ b/src/backend/commands/tablecmds.c
@@ -529,6 +529,8 @@ static void ATExecGenericOptions(Relation rel, List *options);
 static void ATExecEnableRowSecurity(Relation rel);
 static void ATExecDisableRowSecurity(Relation rel);
 static void ATExecForceNoForceRowSecurity(Relation rel, bool force_rls);
+static ObjectAddress ATExecSetCompression(AlteredTableInfo *tab, Relation rel,
+					 const char *column, Node *newValue, LOCKMODE lockmode);
 
 static void index_copy_data(Relation rel, RelFileNode newrnode);
 static const char *storage_name(char c);
@@ -3860,6 +3862,7 @@ AlterTableGetLockLevel(List *cmds)
 				 */
 			case AT_GenericOptions:
 			case AT_AlterColumnGenericOptions:
+			case AT_SetCompression:
 				cmd_lockmode = AccessExclusiveLock;
 				break;
 
@@ -4387,7 +4390,8 @@ ATPrepCmd(List **wqueue, Relation rel, AlterTableCmd *cmd,
 		case AT_DisableRowSecurity:
 		case AT_ForceRowSecurity:
 		case AT_NoForceRowSecurity:
-			ATSimplePermissions(rel, ATT_TABLE);
+		case AT_SetCompression:
+			ATSimplePermissions(rel, ATT_TABLE | ATT_MATVIEW);
 			/* These commands never recurse */
 			/* No command-specific prep needed */
 			pass = AT_PASS_MISC;
@@ -4795,6 +4799,10 @@ ATExecCmd(List **wqueue, AlteredTableInfo *tab, Relation rel,
 			Assert(rel->rd_rel->relkind == RELKIND_INDEX);
 			ATExecAlterCollationRefreshVersion(rel, cmd->object);
 			break;
+		case AT_SetCompression:
+			address = ATExecSetCompression(tab, rel, cmd->name, cmd->def,
+										   lockmode);
+			break;
 		default:				/* oops */
 			elog(ERROR, "unrecognized alter table type: %d",
 				 (int) cmd->subtype);
@@ -5419,6 +5427,9 @@ ATRewriteTable(AlteredTableInfo *tab, Oid OIDNewHeap, LOCKMODE lockmode)
 				slot_getallattrs(oldslot);
 				ExecClearTuple(newslot);
 
+				if (tab->rewrite & AT_REWRITE_ALTER_COMPRESSION)
+					CompareCompressionMethodAndDecompress(oldslot, newTupDesc);
+
 				/* copy attributes */
 				memcpy(newslot->tts_values, oldslot->tts_values,
 					   sizeof(Datum) * oldslot->tts_nvalid);
@@ -7659,6 +7670,71 @@ ATExecSetOptions(Relation rel, const char *colName, Node *options,
 	return address;
 }
 
+/*
+ * Helper function for ATExecSetStorage and ATExecSetCompression
+ *
+ * Set the attcompression or attstorage for the respective index attribute.  If
+ * attcompression is a valid oid then it will update the attcompression for the
+ * index attribute otherwise it will update the attstorage.
+ */
+static void
+ApplyChangesToIndexes(Relation rel, Relation attrelation, AttrNumber attnum,
+					  Oid newcompression, char newstorage, LOCKMODE lockmode)
+{
+	HeapTuple	tuple;
+	ListCell   *lc;
+	Form_pg_attribute attrtuple;
+
+	foreach(lc, RelationGetIndexList(rel))
+	{
+		Oid			indexoid = lfirst_oid(lc);
+		Relation	indrel;
+		AttrNumber	indattnum = 0;
+
+		indrel = index_open(indexoid, lockmode);
+
+		for (int i = 0; i < indrel->rd_index->indnatts; i++)
+		{
+			if (indrel->rd_index->indkey.values[i] == attnum)
+			{
+				indattnum = i + 1;
+				break;
+			}
+		}
+
+		if (indattnum == 0)
+		{
+			index_close(indrel, lockmode);
+			continue;
+		}
+
+		tuple = SearchSysCacheCopyAttNum(RelationGetRelid(indrel), indattnum);
+
+		if (HeapTupleIsValid(tuple))
+		{
+			attrtuple = (Form_pg_attribute) GETSTRUCT(tuple);
+
+			/*
+			 * If a valid compression method Oid is passed then update the
+			 * attcompression, otherwise update the attstorage.
+			 */
+			if (OidIsValid(newcompression))
+				attrtuple->attcompression = newcompression;
+			else
+				attrtuple->attstorage = newstorage;
+
+			CatalogTupleUpdate(attrelation, &tuple->t_self, tuple);
+
+			InvokeObjectPostAlterHook(RelationRelationId,
+									  RelationGetRelid(rel),
+									  attrtuple->attnum);
+
+			heap_freetuple(tuple);
+		}
+
+		index_close(indrel, lockmode);
+	}
+}
 /*
  * ALTER TABLE ALTER COLUMN SET STORAGE
  *
@@ -7674,7 +7750,6 @@ ATExecSetStorage(Relation rel, const char *colName, Node *newValue, LOCKMODE loc
 	Form_pg_attribute attrtuple;
 	AttrNumber	attnum;
 	ObjectAddress address;
-	ListCell   *lc;
 
 	Assert(IsA(newValue, String));
 	storagemode = strVal(newValue);
@@ -7738,47 +7813,8 @@ ATExecSetStorage(Relation rel, const char *colName, Node *newValue, LOCKMODE loc
 	 * Apply the change to indexes as well (only for simple index columns,
 	 * matching behavior of index.c ConstructTupleDescriptor()).
 	 */
-	foreach(lc, RelationGetIndexList(rel))
-	{
-		Oid			indexoid = lfirst_oid(lc);
-		Relation	indrel;
-		AttrNumber	indattnum = 0;
-
-		indrel = index_open(indexoid, lockmode);
-
-		for (int i = 0; i < indrel->rd_index->indnatts; i++)
-		{
-			if (indrel->rd_index->indkey.values[i] == attnum)
-			{
-				indattnum = i + 1;
-				break;
-			}
-		}
-
-		if (indattnum == 0)
-		{
-			index_close(indrel, lockmode);
-			continue;
-		}
-
-		tuple = SearchSysCacheCopyAttNum(RelationGetRelid(indrel), indattnum);
-
-		if (HeapTupleIsValid(tuple))
-		{
-			attrtuple = (Form_pg_attribute) GETSTRUCT(tuple);
-			attrtuple->attstorage = newstorage;
-
-			CatalogTupleUpdate(attrelation, &tuple->t_self, tuple);
-
-			InvokeObjectPostAlterHook(RelationRelationId,
-									  RelationGetRelid(rel),
-									  attrtuple->attnum);
-
-			heap_freetuple(tuple);
-		}
-
-		index_close(indrel, lockmode);
-	}
+	ApplyChangesToIndexes(rel, attrelation, attnum, InvalidOid, newstorage,
+						  lockmode);
 
 	table_close(attrelation, RowExclusiveLock);
 
@@ -14999,6 +15035,92 @@ ATExecGenericOptions(Relation rel, List *options)
 	heap_freetuple(tuple);
 }
 
+/*
+ * ALTER TABLE ALTER COLUMN SET COMPRESSION
+ *
+ * Return value is the address of the modified column
+ */
+static ObjectAddress
+ATExecSetCompression(AlteredTableInfo *tab,
+					 Relation rel,
+					 const char *column,
+					 Node *newValue,
+					 LOCKMODE lockmode)
+{
+	Relation	attrel;
+	HeapTuple	tuple;
+	Form_pg_attribute atttableform;
+	AttrNumber	attnum;
+	char	   *compression;
+	char		typstorage;
+	Oid			cmoid;
+	Datum		values[Natts_pg_attribute];
+	bool		nulls[Natts_pg_attribute];
+	bool		replace[Natts_pg_attribute];
+	ObjectAddress address;
+
+	Assert(IsA(newValue, String));
+	compression = strVal(newValue);
+
+	attrel = table_open(AttributeRelationId, RowExclusiveLock);
+
+	tuple = SearchSysCacheAttName(RelationGetRelid(rel), column);
+	if (!HeapTupleIsValid(tuple))
+		ereport(ERROR,
+				(errcode(ERRCODE_UNDEFINED_COLUMN),
+				 errmsg("column \"%s\" of relation \"%s\" does not exist",
+						column, RelationGetRelationName(rel))));
+
+	/* Prevent them from altering a system attribute */
+	atttableform = (Form_pg_attribute) GETSTRUCT(tuple);
+	attnum = atttableform->attnum;
+	if (attnum <= 0)
+		ereport(ERROR,
+				(errcode(ERRCODE_FEATURE_NOT_SUPPORTED),
+				 errmsg("cannot alter system column \"%s\"", column)));
+
+	typstorage = get_typstorage(atttableform->atttypid);
+
+	/* Prevent from setting compression methods for uncompressible type */
+	if (!IsStorageCompressible(typstorage))
+		ereport(ERROR,
+				(errcode(ERRCODE_FEATURE_NOT_SUPPORTED),
+					errmsg("column data type %s does not support compression",
+						format_type_be(atttableform->atttypid))));
+
+	/* Initialize buffers for new tuple values */
+	memset(values, 0, sizeof(values));
+	memset(nulls, false, sizeof(nulls));
+	memset(replace, false, sizeof(replace));
+
+	/* Get the attribute compression method. */
+	cmoid = GetAttributeCompression(atttableform, compression);
+
+	if (atttableform->attcompression != cmoid)
+		tab->rewrite |= AT_REWRITE_ALTER_COMPRESSION;
+
+	atttableform->attcompression = cmoid;
+	CatalogTupleUpdate(attrel, &tuple->t_self, tuple);
+
+	InvokeObjectPostAlterHook(RelationRelationId,
+							  RelationGetRelid(rel),
+							  atttableform->attnum);
+
+	ReleaseSysCache(tuple);
+
+	/* apply changes to the index column as well */
+	ApplyChangesToIndexes(rel, attrel, attnum, cmoid, '\0', lockmode);
+	table_close(attrel, RowExclusiveLock);
+
+	/* make changes visible */
+	CommandCounterIncrement();
+
+	ObjectAddressSubSet(address, RelationRelationId,
+						RelationGetRelid(rel), atttableform->attnum);
+	return address;
+}
+
+
 /*
  * Preparation phase for SET LOGGED/UNLOGGED
  *
diff --git a/src/backend/parser/gram.y b/src/backend/parser/gram.y
index 39467aa2e0..9683062216 100644
--- a/src/backend/parser/gram.y
+++ b/src/backend/parser/gram.y
@@ -2257,6 +2257,15 @@ alter_table_cmd:
 					n->missing_ok = true;
 					$$ = (Node *)n;
 				}
+			/* ALTER TABLE <name> ALTER [COLUMN] <colname> SET (COMPRESSION <cm>) */
+			| ALTER opt_column ColId SET optColumnCompression
+				{
+					AlterTableCmd *n = makeNode(AlterTableCmd);
+					n->subtype = AT_SetCompression;
+					n->name = $3;
+					n->def = (Node *) makeString($5);;
+					$$ = (Node *)n;
+				}
 			/* ALTER TABLE <name> DROP [COLUMN] IF EXISTS <colname> [RESTRICT|CASCADE] */
 			| DROP opt_column IF_P EXISTS ColId opt_drop_behavior
 				{
diff --git a/src/bin/psql/tab-complete.c b/src/bin/psql/tab-complete.c
index 3a43c09bf6..1da16ebc0a 100644
--- a/src/bin/psql/tab-complete.c
+++ b/src/bin/psql/tab-complete.c
@@ -2093,7 +2093,7 @@ psql_completion(const char *text, int start, int end)
 	/* ALTER TABLE ALTER [COLUMN] <foo> SET */
 	else if (Matches("ALTER", "TABLE", MatchAny, "ALTER", "COLUMN", MatchAny, "SET") ||
 			 Matches("ALTER", "TABLE", MatchAny, "ALTER", MatchAny, "SET"))
-		COMPLETE_WITH("(", "DEFAULT", "NOT NULL", "STATISTICS", "STORAGE");
+		COMPLETE_WITH("(", "COMPRESSION", "DEFAULT", "NOT NULL", "STATISTICS", "STORAGE");
 	/* ALTER TABLE ALTER [COLUMN] <foo> SET ( */
 	else if (Matches("ALTER", "TABLE", MatchAny, "ALTER", "COLUMN", MatchAny, "SET", "(") ||
 			 Matches("ALTER", "TABLE", MatchAny, "ALTER", MatchAny, "SET", "("))
diff --git a/src/include/commands/event_trigger.h b/src/include/commands/event_trigger.h
index 407fd6a978..1800fd8d0c 100644
--- a/src/include/commands/event_trigger.h
+++ b/src/include/commands/event_trigger.h
@@ -32,6 +32,7 @@ typedef struct EventTriggerData
 #define AT_REWRITE_ALTER_PERSISTENCE	0x01
 #define AT_REWRITE_DEFAULT_VAL			0x02
 #define AT_REWRITE_COLUMN_REWRITE		0x04
+#define AT_REWRITE_ALTER_COMPRESSION	0x08
 
 /*
  * EventTriggerData is the node type that is passed as fmgr "context" info
diff --git a/src/include/nodes/parsenodes.h b/src/include/nodes/parsenodes.h
index 78b7f76215..af79aa6534 100644
--- a/src/include/nodes/parsenodes.h
+++ b/src/include/nodes/parsenodes.h
@@ -1856,7 +1856,8 @@ typedef enum AlterTableType
 	AT_AddIdentity,				/* ADD IDENTITY */
 	AT_SetIdentity,				/* SET identity column options */
 	AT_DropIdentity,			/* DROP IDENTITY */
-	AT_AlterCollationRefreshVersion /* ALTER COLLATION ... REFRESH VERSION */
+	AT_AlterCollationRefreshVersion, /* ALTER COLLATION ... REFRESH VERSION */
+	AT_SetCompression			/* SET COMPRESSION */
 } AlterTableType;
 
 typedef struct ReplicaIdentityStmt
diff --git a/src/test/regress/expected/compression.out b/src/test/regress/expected/compression.out
index 2f53c0e489..3f47e5b576 100644
--- a/src/test/regress/expected/compression.out
+++ b/src/test/regress/expected/compression.out
@@ -143,6 +143,65 @@ CREATE TABLE cminh(f1 TEXT COMPRESSION lz4) INHERITS(cmdata);
 NOTICE:  merging column "f1" with inherited definition
 ERROR:  column "f1" has a compression method conflict
 DETAIL:  pglz versus lz4
+-- test alter compression method with rewrite
+ALTER TABLE cmdata ALTER COLUMN f1 SET COMPRESSION lz4;
+\d+ cmdata
+                                        Table "public.cmdata"
+ Column | Type | Collation | Nullable | Default | Storage  | Compression | Stats target | Description 
+--------+------+-----------+----------+---------+----------+-------------+--------------+-------------
+ f1     | text |           |          |         | extended | lz4         |              | 
+Indexes:
+    "idx" btree (f1)
+
+SELECT pg_column_compression(f1) FROM cmdata;
+ pg_column_compression 
+-----------------------
+ lz4
+(1 row)
+
+-- test alter compression method for the materialized view
+ALTER TABLE cmdata1 ALTER COLUMN f1 SET COMPRESSION pglz;
+ALTER MATERIALIZED VIEW mv ALTER COLUMN x SET COMPRESSION lz4;
+REFRESH MATERIALIZED VIEW mv;
+\d+ mv
+                                    Materialized view "public.mv"
+ Column | Type | Collation | Nullable | Default | Storage  | Compression | Stats target | Description 
+--------+------+-----------+----------+---------+----------+-------------+--------------+-------------
+ x      | text |           |          |         | extended | lz4         |              | 
+View definition:
+ SELECT cmdata1.f1 AS x
+   FROM cmdata1;
+
+SELECT pg_column_compression(f1) FROM cmdata1;
+ pg_column_compression 
+-----------------------
+ pglz
+(1 row)
+
+SELECT pg_column_compression(x) FROM mv;
+ pg_column_compression 
+-----------------------
+ lz4
+(1 row)
+
+-- test alter compression method for the partioned table
+ALTER TABLE cmpart ALTER COLUMN f1 SET COMPRESSION pglz;
+SELECT pg_column_compression(f1) FROM cmpart;
+ pg_column_compression 
+-----------------------
+ lz4
+ pglz
+(2 rows)
+
+ALTER TABLE cmpart1 ALTER COLUMN f1 SET COMPRESSION pglz;
+ALTER TABLE cmpart2 ALTER COLUMN f1 SET COMPRESSION lz4;
+SELECT pg_column_compression(f1) FROM cmpart;
+ pg_column_compression 
+-----------------------
+ pglz
+ lz4
+(2 rows)
+
 -- check data is ok
 SELECT length(f1) FROM cmdata;
  length 
diff --git a/src/test/regress/sql/compression.sql b/src/test/regress/sql/compression.sql
index 292c7258ba..bdb85a650b 100644
--- a/src/test/regress/sql/compression.sql
+++ b/src/test/regress/sql/compression.sql
@@ -66,6 +66,27 @@ SELECT pg_column_compression(f1) FROM cmpart;
 CREATE TABLE cminh() INHERITS(cmdata, cmdata1);
 CREATE TABLE cminh(f1 TEXT COMPRESSION lz4) INHERITS(cmdata);
 
+-- test alter compression method with rewrite
+ALTER TABLE cmdata ALTER COLUMN f1 SET COMPRESSION lz4;
+\d+ cmdata
+SELECT pg_column_compression(f1) FROM cmdata;
+
+-- test alter compression method for the materialized view
+ALTER TABLE cmdata1 ALTER COLUMN f1 SET COMPRESSION pglz;
+ALTER MATERIALIZED VIEW mv ALTER COLUMN x SET COMPRESSION lz4;
+REFRESH MATERIALIZED VIEW mv;
+\d+ mv
+SELECT pg_column_compression(f1) FROM cmdata1;
+SELECT pg_column_compression(x) FROM mv;
+
+-- test alter compression method for the partioned table
+ALTER TABLE cmpart ALTER COLUMN f1 SET COMPRESSION pglz;
+SELECT pg_column_compression(f1) FROM cmpart;
+
+ALTER TABLE cmpart1 ALTER COLUMN f1 SET COMPRESSION pglz;
+ALTER TABLE cmpart2 ALTER COLUMN f1 SET COMPRESSION lz4;
+SELECT pg_column_compression(f1) FROM cmpart;
+
 -- check data is ok
 SELECT length(f1) FROM cmdata;
 SELECT length(f1) FROM cmdata1;
-- 
2.23.0