Re: index ops for _int4 and trees?

From: Lincoln Yeoh <lyeoh(at)pop(dot)jaring(dot)my>
To: "Gyozo Papp" <pgerzson(at)freestart(dot)hu>, "PostgreSQL-General" <pgsql-general(at)postgresql(dot)org>
Cc: pgsql-hackers(at)postgresql(dot)org
Subject: Re: index ops for _int4 and trees?
Date: 2001-05-16 01:41:47
Message-ID: 3.0.5.32.20010516094147.0101a400@192.228.128.13
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-general pgsql-hackers

At 06:29 PM 15-05-2001 +0200, Gyozo Papp wrote:
>Oh,
>I forgot that without any contrib you can select rows whose papth to the
root {1,2,3}:
>
>select * from test where lineage[1:3] = '{1,2,3}';
>but the gist indexing (intarray) performs a significant speed increase.

Thanks.

>BTW, lineage represents egde-list of a directed graph or a tree ?
>> insert into test (id,lineage) values ('8','{1,2,3}');
>------------------------------------------------|-^
>> insert into test (id,lineage) values ('9','{1,3,7}');
>It seems to me that node 3 can be accessed from both node 1 and node 3
directly, or it's just a mistake?

It's a mistake in my example.

Aside but related:
Oleg Bartunov also mentioned that subset searches are possible with gist:

select * from table <TABLE> where <array_field> @ '{1,2,3}'

So I've asked him whether his work on gist indexing int arrays can be used
to do substring indexing on text, as a built-in to Postgresql.

I think it can be done. Then subtext_ops here we come :).

If not I'll resort to converting text characters to their code values and
stuffing them into int arrays. Ugly :). Not sure what happens when the
arrays get large.

Cheerio,
Link.

In response to

Browse pgsql-general by date

  From Date Subject
Next Message Al Frick 2001-05-16 02:05:36 Which OS to Use for Postgresql
Previous Message John Clark L. Naldoza 2001-05-16 00:44:16 Re: Not a PG question: SCSI question

Browse pgsql-hackers by date

  From Date Subject
Next Message Steve O'Hagan 2001-05-16 02:18:24 Re: Re: Internet is putting lot of competition fire & heat under Microsoft SQL Server
Previous Message John Clark L. Naldoza 2001-05-16 00:44:16 Re: Not a PG question: SCSI question