[HACKERS] Oddness 7.3 vs 7.4

[Christopher Kings-Lynne]

I notice this in 7.3.4:

test=# create table test (a int4, b int4);
CREATE TABLE
test=# create index idx on test(a) where b is null;
CREATE INDEX
test=# \d test
 Table public.test
 Column |  Type   | Modifiers
+-+---
 a  | integer |
 b  | integer |
Indexes: idx btree (a) WHERE (b IS NULL)
test=# alter table test drop column b;
ALTER TABLE
test=# \d test
 Table public.test
 Column |  Type   | Modifiers
+-+---
 a  | integer |
Indexes: idx btree (a) WHERE (pg.dropped.2 IS NULL)
And this in 7.5CVS:

test=# create table test (a int4, b int4);
CREATE TABLE
test=# create index idx on test(a) where b is null;
CREATE INDEX
test=# alter table test drop column b;
ALTER TABLE
test=# \d test
 Table public.test
 Column |  Type   | Modifiers
+-+---
 a  | integer |
I cannot think of any commit that fixed that???  Is it fixed in 7.3.5?

Chris

---(end of broadcast)---
TIP 8: explain analyze is your friend

Re: [HACKERS] Oddness 7.3 vs 7.4

[Tom Lane]

Christopher Kings-Lynne [EMAIL PROTECTED] writes:
 I cannot think of any commit that fixed that???  Is it fixed in 7.3.5?

7.3.5 exhibits the same misbehavior.  The bug is actually that the index
doesn't get dropped when you drop the column, and it was fixed here:

2003-05-28 12:03  tgl

Replace functional-index facility with expressional indexes.
...
Along the way, ... fix 7.3 oversight
that didn't record dependencies of predicate expressions.

regards, tom lane

---(end of broadcast)---
TIP 9: the planner will ignore your desire to choose an index scan if your
  joining column's datatypes do not match