Re: [HACKERS] WIP: SP-GiST, Space-Partitioned GiST

2011-12-17 Thread Tom Lane
Teodor Sigaev teo...@sigaev.ru writes: [ spgist patch ] I've applied this after a lot of revisions, some cosmetic (better comments etc), some not so much (less bogus vacuum and WAL handling). There are still a number of loose ends that need to be worked on: * The question of whether to store

Re: [HACKERS] WIP: SP-GiST, Space-Partitioned GiST

2011-12-16 Thread Greg Smith
Since this was marked WIP and Tom has now kicked off two side discussions, what I've done is tagged references to each of them as comments to the main patch, then marked this as returned with feedback. Surely what I do in the CF app isn't going to influence what Tom wants to work on, so I'll

Re: [HACKERS] WIP: SP-GiST, Space-Partitioned GiST

2011-12-13 Thread Teodor Sigaev
I wrote: ... the leaf tuple datatype is hard-wired to be After another day's worth of hacking, I now understand the reason for the above: when an index is less than a page and an incoming value would still fit on the root page, the incoming value is simply dumped into a leaf tuple without ever

Re: [HACKERS] WIP: SP-GiST, Space-Partitioned GiST

2011-12-13 Thread Tom Lane
Teodor Sigaev teo...@sigaev.ru writes: However, I now have another question: what is the point of the allTheSame mechanism? It seems to add quite a great deal of complexity, I thought about two options: separate code path in core to support a-lot-of-the-same-values with minimal support in

Re: [HACKERS] WIP: SP-GiST, Space-Partitioned GiST

2011-12-10 Thread Tom Lane
I wrote: ... the leaf tuple datatype is hard-wired to be the same as the indexed column's type. Why is that? It seems to me to be both confusing and restrictive. For instance, if you'd designed the suffix tree opclass just a little differently, it would be wanting to store char not text in

Re: [HACKERS] WIP: SP-GiST, Space-Partitioned GiST

2011-12-09 Thread Tom Lane
Teodor Sigaev teo...@sigaev.ru writes: [ spgist patch ] I've been working through this patch and fixing assorted things. There are a couple of issues that require some discussion: 1. It took me awhile to realize it, but there are actually three different datatypes that can be stored in an

Re: [HACKERS] WIP: SP-GiST, Space-Partitioned GiST

2011-12-06 Thread Tom Lane
Oleg Bartunov o...@sai.msu.su writes: There is one annoying problem under MAC OS (Linux, FreeBSD have no problem), we just can't figure out how to find it, since we are not familiar with MAC OS - it fails to restart after 'kill -9' backend, but only if sources were compiled with -O2

Re: [HACKERS] WIP: SP-GiST, Space-Partitioned GiST

2011-10-06 Thread Oleg Bartunov
We are working on the hackers documentation, hope to submit it before my himalaya track. Oleg On Sun, 2 Oct 2011, Tom Lane wrote: Heikki Linnakangas heikki.linnakan...@enterprisedb.com writes: On 06.09.2011 20:34, Oleg Bartunov wrote: Here is the latest spgist patch, which has all planned

Re: [HACKERS] WIP: SP-GiST, Space-Partitioned GiST

2011-10-02 Thread Tom Lane
Heikki Linnakangas heikki.linnakan...@enterprisedb.com writes: On 06.09.2011 20:34, Oleg Bartunov wrote: Here is the latest spgist patch, which has all planned features as well as all overhead, introduced by concurrency and recovery, so performance measurement should be realistic now. I'm

Re: [HACKERS] WIP: SP-GiST, Space-Partitioned GiST

2011-09-22 Thread Heikki Linnakangas
On 06.09.2011 20:34, Oleg Bartunov wrote: Here is the latest spgist patch, which has all planned features as well as all overhead, introduced by concurrency and recovery, so performance measurement should be realistic now. I'm ignoring the text suffix-tree part of this for now, because of the

Re: [HACKERS] WIP: SP-GiST, Space-Partitioned GiST

2011-09-22 Thread Alexander Korotkov
On Thu, Sep 22, 2011 at 2:05 PM, Heikki Linnakangas heikki.linnakan...@enterprisedb.com wrote: Regarding the quadtree, have you compared the performance of that with Alexander's improved split algorithm? I ran some tests using the test harness I still had lying around from the fast GiST index

Re: [HACKERS] WIP: SP-GiST, Space-Partitioned GiST

2011-09-21 Thread Heikki Linnakangas
On 05.09.2011 09:39, Oleg Bartunov wrote: I attached wrong patch in previous message, sorry ! Here is a last version. One little detail caught my eye: In spgSplitNodeAction, you call SpGistGetBuffer() within a critical section. That should be avoided, SpGistGetBuffer() can easily fail if you

Re: [HACKERS] WIP: SP-GiST, Space-Partitioned GiST

2011-09-06 Thread Oleg Bartunov
Here is the latest spgist patch, which has all planned features as well as all overhead, introduced by concurrency and recovery, so performance measurement should be realistic now. Oleg On Mon, 5 Sep 2011, Oleg Bartunov wrote: I attached wrong patch in previous message, sorry ! Here is a

Re: [HACKERS] WIP: SP-GiST, Space-Partitioned GiST

2011-09-06 Thread Andreas Joseph Krogh
On 09/06/2011 07:34 PM, Oleg Bartunov wrote: Here is the latest spgist patch, which has all planned features as well as all overhead, introduced by concurrency and recovery, so performance measurement should be realistic now. Oleg Sorry for not getting the might-be-obvious here, but will

Re: [HACKERS] WIP: SP-GiST, Space-Partitioned GiST

2011-09-06 Thread Alexander Korotkov
On Tue, Sep 6, 2011 at 10:21 PM, Andreas Joseph Krogh andr...@officenet.nowrote: Sorry for not getting the might-be-obvious here, but will this patch bring indexed substring-search to PG? So queries conceptually equal to this will be possible to index: WHERE som_col @@

Re: [HACKERS] WIP: SP-GiST, Space-Partitioned GiST

2011-09-05 Thread Oleg Bartunov
Attached is updated SP-GiST patch, which provides full logging support and fixed several bugs (Thanks ALexander Korotkov for help). Regards, Oleg On Thu, 1 Sep 2011, Oleg Bartunov wrote: This is updates SP-GiST patch, which fixed one bug and replaced test to the locale independent one. On

Re: [HACKERS] WIP: SP-GiST, Space-Partitioned GiST

2011-09-05 Thread Oleg Bartunov
I attached wrong patch in previous message, sorry ! Here is a last version. This is a new WIP of SP-GiST patch, which provides support for: 1. Concurrent vacuum 2. Vacuum logging 3. WAL replay Oleg On Thu, 1 Sep 2011, Oleg Bartunov wrote: This is updates SP-GiST patch, which fixed one bug and

Re: [HACKERS] WIP: SP-GiST, Space-Partitioned GiST

2011-09-01 Thread Oleg Bartunov
This is updates SP-GiST patch, which fixed one bug and replaced test to the locale independent one. On Wed, 31 Aug 2011, Oleg Bartunov wrote: Hi there, attached is our WIP-patch for 9.2 development source tree, which provides implementation of SP-GiST (prototype was presented at PGCon-2011,

Re: [HACKERS] WIP: SP-GiST, Space-Partitioned GiST

2011-09-01 Thread Alexander Korotkov
Hi! Ie expect some problems in support of comparison operators for text, because locale string comparison can have unexpected behaviour. Let's see the example. Create table with words and add extra leading space to some of them. test=# create table dict(id serial, word text); NOTICE: CREATE