mirrors/postgres
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

Fast default trigger and expand_tuple fixes

Browse Source 
Ensure that triggers get properly filled in tuples for the OLD value.
Also fix the logic of detecting missing null values. The previous logic
failed to detect a missing null column before the first missing column
with a default. Fixing this has simplified the logic a bit.

Regression tests are added to test changes. This should ensure better
coverage of expand_tuple().

Original bug reports, and some code and test scripts from Tomas Vondra

Backpatch to release 11.
This commit is contained in:
Andrew Dunstan 2018-09-24 16:11:24 -04:00
parent bfdd02f88b
commit 9625ab7924
4 changed files with 328 additions and 33 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

51
src/backend/access/common/heaptuple.c
View File

@ -823,44 +823,35 @@ expand_tuple(HeapTuple *targetHeapTuple,
{
if (attrmiss[firstmissingnum].am_present)
break;
else
hasNulls = true;
}
/*
* If there are no more missing values everything else must be NULL
* Now walk the missing attributes. If there is a missing value
* make space for it. Otherwise, it's going to be NULL.
*/
if (firstmissingnum >= natts)
for (attnum = firstmissingnum;
attnum < natts;
attnum++)
{
hasNulls = true;
}
else
{
/*
* Now walk the missing attributes. If there is a missing value
* make space for it. Otherwise, it's going to be NULL.
*/
for (attnum = firstmissingnum;
attnum < natts;
attnum++)
if (attrmiss[attnum].am_present)
{
if (attrmiss[attnum].am_present)
{
Form_pg_attribute att = TupleDescAttr(tupleDesc, attnum);
Form_pg_attribute att = TupleDescAttr(tupleDesc, attnum);
targetDataLen = att_align_datum(targetDataLen,
att->attalign,
att->attlen,
attrmiss[attnum].am_value);
targetDataLen = att_align_datum(targetDataLen,
att->attalign,
att->attlen,
attrmiss[attnum].am_value);
targetDataLen = att_addlength_pointer(targetDataLen,
att->attlen,
attrmiss[attnum].am_value);
}
else
{
/* no missing value, so it must be null */
hasNulls = true;
}
targetDataLen = att_addlength_pointer(targetDataLen,
att->attlen,
attrmiss[attnum].am_value);
}
else
{
/* no missing value, so it must be null */
hasNulls = true;
}
}
} /* end if have missing values */

5
src/backend/commands/trigger.c
View File

@ -3396,7 +3396,10 @@ ltrmark:;
LockBuffer(buffer, BUFFER_LOCK_UNLOCK);
}
result = heap_copytuple(&tuple);
if (HeapTupleHeaderGetNatts(tuple.t_data) < relation->rd_att->natts)
result = heap_expand_tuple(&tuple, relation->rd_att);
else
result = heap_copytuple(&tuple);
ReleaseBuffer(buffer);
return result;

191
src/test/regress/expected/fast_default.out
View File

@ -539,8 +539,197 @@ FROM t1;
1 | 0
(20 rows)
DROP TABLE t1;
DROP TABLE T;
-- test that we account for missing columns without defaults correctly
-- in expand_tuple, and that rows are correctly expanded for triggers
CREATE FUNCTION test_trigger()
RETURNS trigger
LANGUAGE plpgsql
AS $$
begin
raise notice 'old tuple: %', to_json(OLD)::text;
if TG_OP = 'DELETE'
then
return OLD;
else
return NEW;
end if;
end;
$$;
-- 2 new columns, both have defaults
CREATE TABLE t (id serial PRIMARY KEY, a int, b int, c int);
INSERT INTO t (a,b,c) VALUES (1,2,3);
ALTER TABLE t ADD COLUMN x int NOT NULL DEFAULT 4;
ALTER TABLE t ADD COLUMN y int NOT NULL DEFAULT 5;
CREATE TRIGGER a BEFORE UPDATE ON t FOR EACH ROW EXECUTE PROCEDURE test_trigger();
SELECT * FROM t;
id | a | b | c | x | y
----+---+---+---+---+---
1 | 1 | 2 | 3 | 4 | 5
(1 row)
UPDATE t SET y = 2;
NOTICE: old tuple: {"id":1,"a":1,"b":2,"c":3,"x":4,"y":5}
SELECT * FROM t;
id | a | b | c | x | y
----+---+---+---+---+---
1 | 1 | 2 | 3 | 4 | 2
(1 row)
DROP TABLE t;
-- 2 new columns, first has default
CREATE TABLE t (id serial PRIMARY KEY, a int, b int, c int);
INSERT INTO t (a,b,c) VALUES (1,2,3);
ALTER TABLE t ADD COLUMN x int NOT NULL DEFAULT 4;
ALTER TABLE t ADD COLUMN y int;
CREATE TRIGGER a BEFORE UPDATE ON t FOR EACH ROW EXECUTE PROCEDURE test_trigger();
SELECT * FROM t;
id | a | b | c | x | y
----+---+---+---+---+---
1 | 1 | 2 | 3 | 4 |
(1 row)
UPDATE t SET y = 2;
NOTICE: old tuple: {"id":1,"a":1,"b":2,"c":3,"x":4,"y":null}
SELECT * FROM t;
id | a | b | c | x | y
----+---+---+---+---+---
1 | 1 | 2 | 3 | 4 | 2
(1 row)
DROP TABLE t;
-- 2 new columns, second has default
CREATE TABLE t (id serial PRIMARY KEY, a int, b int, c int);
INSERT INTO t (a,b,c) VALUES (1,2,3);
ALTER TABLE t ADD COLUMN x int;
ALTER TABLE t ADD COLUMN y int NOT NULL DEFAULT 5;
CREATE TRIGGER a BEFORE UPDATE ON t FOR EACH ROW EXECUTE PROCEDURE test_trigger();
SELECT * FROM t;
id | a | b | c | x | y
----+---+---+---+---+---
1 | 1 | 2 | 3 | | 5
(1 row)
UPDATE t SET y = 2;
NOTICE: old tuple: {"id":1,"a":1,"b":2,"c":3,"x":null,"y":5}
SELECT * FROM t;
id | a | b | c | x | y
----+---+---+---+---+---
1 | 1 | 2 | 3 | | 2
(1 row)
DROP TABLE t;
-- 2 new columns, neither has default
CREATE TABLE t (id serial PRIMARY KEY, a int, b int, c int);
INSERT INTO t (a,b,c) VALUES (1,2,3);
ALTER TABLE t ADD COLUMN x int;
ALTER TABLE t ADD COLUMN y int;
CREATE TRIGGER a BEFORE UPDATE ON t FOR EACH ROW EXECUTE PROCEDURE test_trigger();
SELECT * FROM t;
id | a | b | c | x | y
----+---+---+---+---+---
1 | 1 | 2 | 3 | |
(1 row)
UPDATE t SET y = 2;
NOTICE: old tuple: {"id":1,"a":1,"b":2,"c":3,"x":null,"y":null}
SELECT * FROM t;
id | a | b | c | x | y
----+---+---+---+---+---
1 | 1 | 2 | 3 | | 2
(1 row)
DROP TABLE t;
-- same as last 4 tests but here the last original column has a NULL value
-- 2 new columns, both have defaults
CREATE TABLE t (id serial PRIMARY KEY, a int, b int, c int);
INSERT INTO t (a,b,c) VALUES (1,2,NULL);
ALTER TABLE t ADD COLUMN x int NOT NULL DEFAULT 4;
ALTER TABLE t ADD COLUMN y int NOT NULL DEFAULT 5;
CREATE TRIGGER a BEFORE UPDATE ON t FOR EACH ROW EXECUTE PROCEDURE test_trigger();
SELECT * FROM t;
id | a | b | c | x | y
----+---+---+---+---+---
1 | 1 | 2 | | 4 | 5
(1 row)
UPDATE t SET y = 2;
NOTICE: old tuple: {"id":1,"a":1,"b":2,"c":null,"x":4,"y":5}
SELECT * FROM t;
id | a | b | c | x | y
----+---+---+---+---+---
1 | 1 | 2 | | 4 | 2
(1 row)
DROP TABLE t;
-- 2 new columns, first has default
CREATE TABLE t (id serial PRIMARY KEY, a int, b int, c int);
INSERT INTO t (a,b,c) VALUES (1,2,NULL);
ALTER TABLE t ADD COLUMN x int NOT NULL DEFAULT 4;
ALTER TABLE t ADD COLUMN y int;
CREATE TRIGGER a BEFORE UPDATE ON t FOR EACH ROW EXECUTE PROCEDURE test_trigger();
SELECT * FROM t;
id | a | b | c | x | y
----+---+---+---+---+---
1 | 1 | 2 | | 4 |
(1 row)
UPDATE t SET y = 2;
NOTICE: old tuple: {"id":1,"a":1,"b":2,"c":null,"x":4,"y":null}
SELECT * FROM t;
id | a | b | c | x | y
----+---+---+---+---+---
1 | 1 | 2 | | 4 | 2
(1 row)
DROP TABLE t;
-- 2 new columns, second has default
CREATE TABLE t (id serial PRIMARY KEY, a int, b int, c int);
INSERT INTO t (a,b,c) VALUES (1,2,NULL);
ALTER TABLE t ADD COLUMN x int;
ALTER TABLE t ADD COLUMN y int NOT NULL DEFAULT 5;
CREATE TRIGGER a BEFORE UPDATE ON t FOR EACH ROW EXECUTE PROCEDURE test_trigger();
SELECT * FROM t;
id | a | b | c | x | y
----+---+---+---+---+---
1 | 1 | 2 | | | 5
(1 row)
UPDATE t SET y = 2;
NOTICE: old tuple: {"id":1,"a":1,"b":2,"c":null,"x":null,"y":5}
SELECT * FROM t;
id | a | b | c | x | y
----+---+---+---+---+---
1 | 1 | 2 | | | 2
(1 row)
DROP TABLE t;
-- 2 new columns, neither has default
CREATE TABLE t (id serial PRIMARY KEY, a int, b int, c int);
INSERT INTO t (a,b,c) VALUES (1,2,NULL);
ALTER TABLE t ADD COLUMN x int;
ALTER TABLE t ADD COLUMN y int;
CREATE TRIGGER a BEFORE UPDATE ON t FOR EACH ROW EXECUTE PROCEDURE test_trigger();
SELECT * FROM t;
id | a | b | c | x | y
----+---+---+---+---+---
1 | 1 | 2 | | |
(1 row)
UPDATE t SET y = 2;
NOTICE: old tuple: {"id":1,"a":1,"b":2,"c":null,"x":null,"y":null}
SELECT * FROM t;
id | a | b | c | x | y
----+---+---+---+---+---
1 | 1 | 2 | | | 2
(1 row)
DROP TABLE t;
-- cleanup
DROP FUNCTION test_trigger();
DROP TABLE t1;
DROP FUNCTION set(name);
DROP FUNCTION comp();
DROP TABLE m;

114
src/test/regress/sql/fast_default.sql
View File

@ -360,8 +360,120 @@ SELECT a,
AS z
FROM t1;
DROP TABLE t1;
DROP TABLE T;
-- test that we account for missing columns without defaults correctly
-- in expand_tuple, and that rows are correctly expanded for triggers
CREATE FUNCTION test_trigger()
RETURNS trigger
LANGUAGE plpgsql
AS $$
begin
raise notice 'old tuple: %', to_json(OLD)::text;
if TG_OP = 'DELETE'
then
return OLD;
else
return NEW;
end if;
end;
$$;
-- 2 new columns, both have defaults
CREATE TABLE t (id serial PRIMARY KEY, a int, b int, c int);
INSERT INTO t (a,b,c) VALUES (1,2,3);
ALTER TABLE t ADD COLUMN x int NOT NULL DEFAULT 4;
ALTER TABLE t ADD COLUMN y int NOT NULL DEFAULT 5;
CREATE TRIGGER a BEFORE UPDATE ON t FOR EACH ROW EXECUTE PROCEDURE test_trigger();
SELECT * FROM t;
UPDATE t SET y = 2;
SELECT * FROM t;
DROP TABLE t;
-- 2 new columns, first has default
CREATE TABLE t (id serial PRIMARY KEY, a int, b int, c int);
INSERT INTO t (a,b,c) VALUES (1,2,3);
ALTER TABLE t ADD COLUMN x int NOT NULL DEFAULT 4;
ALTER TABLE t ADD COLUMN y int;
CREATE TRIGGER a BEFORE UPDATE ON t FOR EACH ROW EXECUTE PROCEDURE test_trigger();
SELECT * FROM t;
UPDATE t SET y = 2;
SELECT * FROM t;
DROP TABLE t;
-- 2 new columns, second has default
CREATE TABLE t (id serial PRIMARY KEY, a int, b int, c int);
INSERT INTO t (a,b,c) VALUES (1,2,3);
ALTER TABLE t ADD COLUMN x int;
ALTER TABLE t ADD COLUMN y int NOT NULL DEFAULT 5;
CREATE TRIGGER a BEFORE UPDATE ON t FOR EACH ROW EXECUTE PROCEDURE test_trigger();
SELECT * FROM t;
UPDATE t SET y = 2;
SELECT * FROM t;
DROP TABLE t;
-- 2 new columns, neither has default
CREATE TABLE t (id serial PRIMARY KEY, a int, b int, c int);
INSERT INTO t (a,b,c) VALUES (1,2,3);
ALTER TABLE t ADD COLUMN x int;
ALTER TABLE t ADD COLUMN y int;
CREATE TRIGGER a BEFORE UPDATE ON t FOR EACH ROW EXECUTE PROCEDURE test_trigger();
SELECT * FROM t;
UPDATE t SET y = 2;
SELECT * FROM t;
DROP TABLE t;
-- same as last 4 tests but here the last original column has a NULL value
-- 2 new columns, both have defaults
CREATE TABLE t (id serial PRIMARY KEY, a int, b int, c int);
INSERT INTO t (a,b,c) VALUES (1,2,NULL);
ALTER TABLE t ADD COLUMN x int NOT NULL DEFAULT 4;
ALTER TABLE t ADD COLUMN y int NOT NULL DEFAULT 5;
CREATE TRIGGER a BEFORE UPDATE ON t FOR EACH ROW EXECUTE PROCEDURE test_trigger();
SELECT * FROM t;
UPDATE t SET y = 2;
SELECT * FROM t;
DROP TABLE t;
-- 2 new columns, first has default
CREATE TABLE t (id serial PRIMARY KEY, a int, b int, c int);
INSERT INTO t (a,b,c) VALUES (1,2,NULL);
ALTER TABLE t ADD COLUMN x int NOT NULL DEFAULT 4;
ALTER TABLE t ADD COLUMN y int;
CREATE TRIGGER a BEFORE UPDATE ON t FOR EACH ROW EXECUTE PROCEDURE test_trigger();
SELECT * FROM t;
UPDATE t SET y = 2;
SELECT * FROM t;
DROP TABLE t;
-- 2 new columns, second has default
CREATE TABLE t (id serial PRIMARY KEY, a int, b int, c int);
INSERT INTO t (a,b,c) VALUES (1,2,NULL);
ALTER TABLE t ADD COLUMN x int;
ALTER TABLE t ADD COLUMN y int NOT NULL DEFAULT 5;
CREATE TRIGGER a BEFORE UPDATE ON t FOR EACH ROW EXECUTE PROCEDURE test_trigger();
SELECT * FROM t;
UPDATE t SET y = 2;
SELECT * FROM t;
DROP TABLE t;
-- 2 new columns, neither has default
CREATE TABLE t (id serial PRIMARY KEY, a int, b int, c int);
INSERT INTO t (a,b,c) VALUES (1,2,NULL);
ALTER TABLE t ADD COLUMN x int;
ALTER TABLE t ADD COLUMN y int;
CREATE TRIGGER a BEFORE UPDATE ON t FOR EACH ROW EXECUTE PROCEDURE test_trigger();
SELECT * FROM t;
UPDATE t SET y = 2;
SELECT * FROM t;
DROP TABLE t;
-- cleanup
DROP FUNCTION test_trigger();
DROP TABLE t1;
DROP FUNCTION set(name);
DROP FUNCTION comp();
DROP TABLE m;