From: | Daniel Newman <dtnewman(at)gmail(dot)com> |
---|---|
To: | Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> |
Cc: | danielnewman(at)umich(dot)edu, pgsql-bugs(at)postgresql(dot)org |
Subject: | Re: BUG #14210: filter by "=" constraint doesn't work when hash index is present on a column |
Date: | 2016-06-23 17:31:32 |
Message-ID: | CALo0FatmGhupZU02+2WwN3v=rDoPuEkY_HhMQRhmfdQxerXA8Q@mail.gmail.com |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Thread: | |
Lists: | pgsql-bugs |
Yes. If i delete the index and recreate it, the bug is replicated. I have
uploaded a pg_dump .sql file to
https://www.dropbox.com/s/4dxuo2uthj721od/hash_issue_db.sql?dl=0 that you
can use to recreate the issue.
After creating the database when I run:
>>> select * from hash_issue_table where hash_issue_column = '1';
I get no results.
When I run:
>>> drop index hash_issue_index;
>>> select * from hash_issue_table where hash_issue_column = '1';
I get 531 rows of results.
When I run:
>>> create index hash_issue_index on hash_issue_table using
hash(hash_issue_column);
>>> select * from hash_issue_table where hash_issue_column = '1';
I get 0 results again. I have repeated this successfully multiple times and
on a coworker's machine as well.
Interestingly, I modified the pg_dump file a bit. At the end, it says:
CREATE INDEX hash_issue_index ON hash_issue_table USING hash
> (hash_issue_column);
>
> DROP INDEX hash_issue_index;
>
> CREATE INDEX hash_issue_index ON hash_issue_table USING hash
> (hash_issue_column);
>
This is because the issue was not replicating (for some reason) when it
built the index the first time.
On Thu, Jun 23, 2016 at 1:04 PM, Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> wrote:
> danielnewman(at)umich(dot)edu writes:
> > When I run the query `select * from hash_issue_table where
> hash_issue_column
> > like '2184';` I get 701 results. When I run the query
> > When I run the query `select * from hash_issue_table where
> hash_issue_column
> > = '2184';`, I get 0 results. However, if I drop the hash index and rerun
> the
> > second query, I get the same results.
>
> Sounds like a corrupted hash index to me; which is not terribly surprising
> given hash indexes' lack of WAL support. Can you reproduce this after
> rebuilding the hash index?
>
> regards, tom lane
>
From | Date | Subject | |
---|---|---|---|
Next Message | Vladimir Borodin | 2016-06-23 17:53:07 | Re: BUG #14189: Corrupted WAL through replication protocol |
Previous Message | Alvaro Herrera | 2016-06-23 17:04:55 | Re: about and feature matrix on site |