Re: pgbench: Skipping the creating primary keys after initialization

From: Masahiko Sawada <sawada(dot)mshk(at)gmail(dot)com>
To: Fabien COELHO <coelho(at)cri(dot)ensmp(dot)fr>
Cc: Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>, Robert Haas <robertmhaas(at)gmail(dot)com>, PostgreSQL-development <pgsql-hackers(at)postgresql(dot)org>
Subject: Re: pgbench: Skipping the creating primary keys after initialization
Date: 2017-08-24 07:12:29
Views: Raw Message | Whole Thread | Download mbox
Lists: pgsql-hackers

On Tue, Aug 22, 2017 at 5:15 PM, Fabien COELHO <coelho(at)cri(dot)ensmp(dot)fr> wrote:
>>> Why not allow -I as a short option for --custom-initialize?
>> Other options for similar purpose such as --foreign-keys also have
>> only a long option. Since I think --custom-initialize option is in the
>> same context as other options I didn't add short option to it for now.
>> Because the options name is found by a prefix searching we can use a
>> short name --cu for now.
> Hmmm. I like short options:-)

Okay, I added -I option for custom initialization :)

>> I'm inclined to remove the restriction so that we can specify
>> --foreign-keys, --no-vacuum and --custom-initialize at the same time.
> Ok. I favor that as well.

If building foreign keys command is not specified in -I but
--foreign-keys options is specified (e.g. pgbench -i -I tpd
--foreign-keys) I think we can add 'f' to the end of the
initialization commands.

>> I think a list of char would be better here rather than a single
>> malloced string to remove particular initialization step easily.
>> Thought?
> My thought is not to bother with a list of char.
> To remove a char from a string, I suggest to allow ' ' to stand for nothing
> and be skipped, so that substituting a letter by space would simply remove
> an initialization phase.

I think we should not add/remove a command of initialization commands
during parsing pgbench options in order to not depend on its order.
Therefore, if -I, --foreign-keys and --no-vacuum are specified at the
same time, what we do is removing some 'v' commands if novacuum and
adding a 'f' command if foreignkey. Also we expect that the length of
initialization steps would not long. Using malloced string would less
the work. Ok, I changed the patch so.

Attached latest v4 patch. Please review it.


Masahiko Sawada
NTT Open Source Software Center

Attachment Content-Type Size
pgbench_custom_initialization_v4.patch application/octet-stream 11.8 KB

In response to


Browse pgsql-hackers by date

  From Date Subject
Next Message Pavel Stehule 2017-08-24 07:20:10 Re: proposal: psql command \graw
Previous Message Torsten Zuehlsdorff 2017-08-24 07:07:03 Re: Thoughts on unit testing?