Re: ctidscan as an example of custom-scan (Re: [HACKERS] [v9.5] Custom Plan API)

2015-07-15 Thread Kouhei Kaigai
of custom-scan (Re: [HACKERS] [v9.5] Custom Plan API) On Wed, Jul 15, 2015 at 10:12 AM, Kouhei Kaigai kai...@ak.jp.nec.com wrote: We never guarantee the interface compatibility between major version up. If we add/modify interface on v9.6, it is duty for developer of extensions to follow the new

Re: ctidscan as an example of custom-scan (Re: [HACKERS] [v9.5] Custom Plan API)

2015-07-15 Thread Michael Paquier
On Wed, Jul 15, 2015 at 10:12 AM, Kouhei Kaigai kai...@ak.jp.nec.com wrote: We never guarantee the interface compatibility between major version up. If we add/modify interface on v9.6, it is duty for developer of extensions to follow the new version, even not specific to custom-scan provider.

Re: ctidscan as an example of custom-scan (Re: [HACKERS] [v9.5] Custom Plan API)

2015-07-15 Thread Robert Haas
On Tue, Jul 14, 2015 at 6:04 PM, Tom Lane t...@sss.pgh.pa.us wrote: Robert Haas robertmh...@gmail.com writes: Both you and Andres have articulated the concern that CustomScan isn't actually useful, but I still don't really understand why not. I'm curious, for example, whether CustomScan would

Re: ctidscan as an example of custom-scan (Re: [HACKERS] [v9.5] Custom Plan API)

2015-07-15 Thread Robert Haas
On Wed, Jul 15, 2015 at 2:28 AM, Michael Paquier michael.paqu...@gmail.com wrote: On Wed, Jul 15, 2015 at 10:12 AM, Kouhei Kaigai kai...@ak.jp.nec.com wrote: We never guarantee the interface compatibility between major version up. If we add/modify interface on v9.6, it is duty for developer of

Re: ctidscan as an example of custom-scan (Re: [HACKERS] [v9.5] Custom Plan API)

2015-07-14 Thread Robert Haas
On Fri, Jul 10, 2015 at 8:55 AM, Heikki Linnakangas hlinn...@iki.fi wrote: If somebody still needs it, I'll rebase and adjust the patch towards the latest custom-scan interface. However, I cannot be motivated for the feature nobody wants. Robert, can you weigh in on this? Do we currently have

Re: ctidscan as an example of custom-scan (Re: [HACKERS] [v9.5] Custom Plan API)

2015-07-14 Thread Robert Haas
On Tue, Jul 14, 2015 at 3:07 PM, Alvaro Herrera alvhe...@2ndquadrant.com wrote: We don't have anything that currently tests the Custom Scan interface in the tree. The question is how important that is, and whether it's worth having what's basically a toy implementation just to demonstrate

Re: ctidscan as an example of custom-scan (Re: [HACKERS] [v9.5] Custom Plan API)

2015-07-14 Thread Tom Lane
Robert Haas robertmh...@gmail.com writes: Both you and Andres have articulated the concern that CustomScan isn't actually useful, but I still don't really understand why not. I'm curious, for example, whether CustomScan would have been sufficient to build TABLESAMPLE, and if not, why not.

Re: ctidscan as an example of custom-scan (Re: [HACKERS] [v9.5] Custom Plan API)

2015-07-14 Thread Robert Haas
On Tue, Jul 14, 2015 at 4:47 PM, Tom Lane t...@sss.pgh.pa.us wrote: I think this ties into my core unhappiness with the customscan stuff, which is that I don't believe it's *possible* to do anything of very great interest with it. I think anything really useful will require core code

Re: ctidscan as an example of custom-scan (Re: [HACKERS] [v9.5] Custom Plan API)

2015-07-14 Thread Alvaro Herrera
Robert Haas wrote: We don't have anything that currently tests the Custom Scan interface in the tree. The question is how important that is, and whether it's worth having what's basically a toy implementation just to demonstrate that the feature can work. If so, I think ctidscan is as good

Re: ctidscan as an example of custom-scan (Re: [HACKERS] [v9.5] Custom Plan API)

2015-07-14 Thread Tom Lane
Robert Haas robertmh...@gmail.com writes: On Tue, Jul 14, 2015 at 3:07 PM, Alvaro Herrera alvhe...@2ndquadrant.com wrote: As a general principle, I think it's a good idea to have a module that's mostly just a skeleton that guides people into writing something real to use whatever API is being

Re: ctidscan as an example of custom-scan (Re: [HACKERS] [v9.5] Custom Plan API)

2015-07-14 Thread Kouhei Kaigai
(Re: [HACKERS] [v9.5] Custom Plan API) Robert Haas robertmh...@gmail.com writes: On Tue, Jul 14, 2015 at 3:07 PM, Alvaro Herrera alvhe...@2ndquadrant.com wrote: As a general principle, I think it's a good idea to have a module that's mostly just a skeleton that guides people

Re: ctidscan as an example of custom-scan (Re: [HACKERS] [v9.5] Custom Plan API)

2015-07-14 Thread Kouhei Kaigai
Or, taking the example of a GpuScan node, it's essentially impossible to persuade the planner to delegate any expensive function calculations, aggregates, etc to such a node; much less teach it that that way is cheaper than doing such things the usual way. So yeah, KaiGai-san may have a

Re: ctidscan as an example of custom-scan (Re: [HACKERS] [v9.5] Custom Plan API)

2015-07-10 Thread Heikki Linnakangas
On 07/02/2015 08:21 AM, Kouhei Kaigai wrote: Folks, Moved this patch to next CF 2015-02 because of lack of review(ers). Do we still need this patch as contrib module? It was originally required it as example of custom-scan interface last summer, however, here was no strong requirement

Re: ctidscan as an example of custom-scan (Re: [HACKERS] [v9.5] Custom Plan API)

2015-07-01 Thread Kouhei Kaigai
- From: pgsql-hackers-ow...@postgresql.org [mailto:pgsql-hackers-ow...@postgresql.org] On Behalf Of Michael Paquier Sent: Friday, February 13, 2015 4:41 PM To: Kaigai Kouhei(海外 浩平) Cc: Jim Nasby; Robert Haas; Simon Riggs; Kohei KaiGai; PgHacker Subject: Re: ctidscan as an example of custom-scan (Re

Re: ctidscan as an example of custom-scan (Re: [HACKERS] [v9.5] Custom Plan API)

2015-02-12 Thread Michael Paquier
On Tue, Jan 6, 2015 at 10:51 PM, Kouhei Kaigai kai...@ak.jp.nec.com wrote: Jim, Thanks for your reviewing the patch. The attached patch is revised one according to your suggestion, and also includes bug fix I could found. * Definitions of TIDOperator was moved to pg_operator.h as

Re: ctidscan as an example of custom-scan (Re: [HACKERS] [v9.5] Custom Plan API)

2014-12-24 Thread Kouhei Kaigai
On Mon, Dec 15, 2014 at 4:55 PM, Kouhei Kaigai kai...@ak.jp.nec.com wrote: I'm not certain whether we should have this functionality in contrib from the perspective of workload that can help, but its major worth is for an example of custom-scan interface. worker_spi is now in

Re: ctidscan as an example of custom-scan (Re: [HACKERS] [v9.5] Custom Plan API)

2014-12-15 Thread Michael Paquier
On Mon, Dec 15, 2014 at 4:55 PM, Kouhei Kaigai kai...@ak.jp.nec.com wrote: I'm not certain whether we should have this functionality in contrib from the perspective of workload that can help, but its major worth is for an example of custom-scan interface. worker_spi is now in src/test/modules.

Re: ctidscan as an example of custom-scan (Re: [HACKERS] [v9.5] Custom Plan API)

2014-12-15 Thread Kouhei Kaigai
On Mon, Dec 15, 2014 at 4:55 PM, Kouhei Kaigai kai...@ak.jp.nec.com wrote: I'm not certain whether we should have this functionality in contrib from the perspective of workload that can help, but its major worth is for an example of custom-scan interface. worker_spi is now in

ctidscan as an example of custom-scan (Re: [HACKERS] [v9.5] Custom Plan API)

2014-12-14 Thread Kouhei Kaigai
This attached patch adds a code example of custom-scan interface. This custom-scan provider (ctidscan) performs almost same as built-in SeqScan plan, but can produce same results with less page scan in case when qualifier of relation scan has inequality operators (, =, or =) on ctid system