pg.ddx.io  pgsql-sql@postgresql.org mailing list archive  

help / color / mirror / Atom feed

query or design question
2+ messages / 2 participants
[nested] [flat]

* query or design question
@ 2003-10-20 10:58  Adam Witney <awitney@sghms.ac.uk>
  0 siblings, 1 reply; 2+ messages in thread

From: Adam Witney @ 2003-10-20 10:58 UTC (permalink / raw)
  To: pgsql-sql

Hi,

I have a table like so

CREATE TABLE imagene (
id             int
bioassay_id    int
gene_id        text
s_row          int
s_column       int
s_meta_row     int
s_meta_column  int
sig_median     numeric
bkg_median     numeric
);

Rows are unique on (bioassay_id, gene_id, s_row, s_column, s_meta_row,
s_meta_column)

They are grouped like this (note, the counts will not always be the same)

cabbage=# select bioassay_id, count(*) from imagene group by bioassay_id;
 bioassay_id | count
-------------+-------
         106 | 10944
         107 | 10944


And I need to generate an output like this......

             bioassay_id=106        bioassay_id=107
         ----------------------  ----------------------
gene_id, sig_median, bkg_median, sig_median, bkg_median



I can do something like this

SELECT a.gene_id,
         a.sig_median, a.bkg_median,
         b.sig_median, b.bkg_median
   FROM 
       imagene a,
       imagene b
   WHERE 
       a.s_meta_row = b.s_meta_row AND
       a.s_meta_col = b.s_meta_col AND
       a.s_row = b.s_row AND
       a.s_column = b.s_column AND
       a.bioassay_id = 106 AND
       b.bioassay_id = 107;


But this is quite slow... And not generic as I may need to generate the
output for more than two bioassay_ids. Also I may need to do an outer join
to make sure I get all rows from both data sets?

I am not sure if a crosstab function will do this?

Any ideas of pointers would be greatly appreciated

Thanks

Adam


-- 
This message has been scanned for viruses and
dangerous content by MailScanner, and is
believed to be clean.




^ permalink  raw  reply  [nested|flat] 2+ messages in thread

* Re: query or design question
@ 2003-10-20 17:12  Richard Huxton <dev@archonet.com>
  parent: Adam Witney <awitney@sghms.ac.uk>
  0 siblings, 0 replies; 2+ messages in thread

From: Richard Huxton @ 2003-10-20 17:12 UTC (permalink / raw)
  To: Adam Witney <awitney@sghms.ac.uk>; pgsql-sql

On Monday 20 October 2003 11:58, Adam Witney wrote:
> Hi,
>
> I have a table like so
>
> CREATE TABLE imagene (
> id             int
> bioassay_id    int
> gene_id        text
> s_row          int
> s_column       int
> s_meta_row     int
> s_meta_column  int
> sig_median     numeric
> bkg_median     numeric
> );

> And I need to generate an output like this......
>
>              bioassay_id=106        bioassay_id=107
>          ----------------------  ----------------------
> gene_id, sig_median, bkg_median, sig_median, bkg_median


> I am not sure if a crosstab function will do this?

There are some examples in contrib/tablefunc that might well be useful for 
you. You should also check out the "Set Returning Functions" article on 
http://techdocs.postgresql.org/

If you find your requirements getting particularly complex, I'd be tempted to 
do the hard work in the application.

-- 
  Richard Huxton
  Archonet Ltd



^ permalink  raw  reply  [nested|flat] 2+ messages in thread

end of thread, other threads:[~2003-10-20 17:12 UTC | newest]

Thread overview: 2+ messages (download: mbox mbox.gz follow: Atom feed)
-- links below jump to the message on this page --
2003-10-20 10:58 query or design question Adam Witney <awitney@sghms.ac.uk>
2003-10-20 17:12 ` Richard Huxton <dev@archonet.com>

This inbox is served by DDX for PostgreSQL; see mirroring instructions
for how to clone and mirror all data and code used for this inbox