Re: "Some tests to cover hash_index"

On Mon, Sep 19, 2016 at 8:44 PM, Mithun Cy <mithun(dot)cy(at)enterprisedb(dot)com> wrote:
> On Sat, Aug 6, 2016 at 9:41 AM, Amit Kapila <amit(dot)kapila16(at)gmail(dot)com> wrote:
>> I wonder why you have included a new file for these tests, why can't be
>> these added to existing hash_index.sql.
> tests in hash_index.sql did not cover overflow pages, above tests were for
> mainly for them.
>

The name of file hash_index_split suggests it focus on split. Are the
tests focussed more on overflow pages or on split of hash index?

So I thought having a separate test file can help
> enabling/disabling them in schedule files, when we do not want them running
> as it take slightly high time. If you think otherwise I will reconsider will
> add tests to hash_index.sql.
>

I think you have a point, but not sure if it is worth to add a
separate file. It might be tricky to choose which file to add new
tests for hash_indexes. Anybody else have opinion on this point?

Can you check how much time it takes as compare to btree or brin index tests?

I am facing below diff with your new patch.

***************
*** 1,4 ****
! --
-- Cause some overflow insert and splits.
--
CREATE TABLE hash_split_heap (keycol INT);
--- 1,4 ----
! --
-- Cause some overflow insert and splits.
--
CREATE TABLE hash_split_heap (keycol INT);

======================================================================

There is an extra space in expected file which is leading to above failure.

--
With Regards,
Amit Kapila.
EnterpriseDB: http://www.enterprisedb.com