Обсуждение: merge two rows where value are null

Hi all,
sorry if OT. I have the following using case:

Name, Surname , Job, Hobby, Pet, address
John, Smith,   NULL, photo, NULL, NULL
John, Smith,   student, NULL, cat, NULL

by using name and surname as selecting key, I want :

John, Smith,   student, photo, cat, NULL

that is I want to substitute the null valus of the LAST row with not-null
value of previous one.

how  can I perform that ? in may installed vesion there aren't function as
merge and collect.

Thanks in advance

GIU

 --
 Email.it, the professional e-mail, gratis per te: http://www.email.it/f

 Sponsor:
 Entra in Mondolastminute, centinaia di offerte ti aspettano per le tue
vacanze a prezzi lastminute!
 Clicca qui: http://adv.email.it/cgi-bin/foclick.cgi?mid=6851&d=20070814

On Tue, 14 Aug 2007, giuseppe.derossi@email.it wrote:

> Name, Surname , Job, Hobby, Pet, address
> John, Smith,   NULL, photo, NULL, NULL
> John, Smith,   student, NULL, cat, NULL

> by using name and surname as selecting key, I want :
>
> John, Smith,   student, photo, cat, NULL

If you are sure there's no conflict (multiple values) for a (name,
   surname) key, one simple solution might be (surely not an efficient one)

select
(select distinct Name from thetable where Name = 'name1' and Surname = 'surname1' where Name is not null),
(select distinct Surname from thetable where Name = 'name1' and Surname = 'surname1' where Surname is not null),
(select distinct Job from thetable where Name = 'name1' and Surname = 'surname1' where Job is not null),
(select distinct Hobby from thetable where Name = 'name1' and Surname = 'surname1' where Hobby is not null),
(select distinct Pet from thetable where Name = 'name1' and Surname = 'surname1' where Pet is not null),
(select distinct address from thetable where Name = 'name1' and Surname = 'surname1' where address is not null)

This will fail if you also have
John, Smith,   student, NULL, dog, NULL

in addition to
> John, Smith,   student, NULL, cat, NULL

You can test for offending rows by:

select
array_to_string( array( select distinct Pet from thetable where Name = 'name1' and Surname = 'surname1' where Pet is
notnull),',') 

and so on...


Regards,

Ben K.
Developer
http://benix.tamu.edu

Thanks for the suggestion,
I was studying a solution which foresees the 'case when' construct...
now I can use it in addition to  array_to_string . In fact my solution
failed too if <it's raining cats and dogs :-) >, but if there are some more
values I get them all as well.

thanks

    --------- Original Message --------
    Da: Ben Kim <bkim@tamu.edu>
    To:
            Cc: pgsql-admin@postgresql.org
    Oggetto: Re: [ADMIN] merge two rows where value are null
    Data: 14/08/07 15:38

    >
>
>
> On Tue, 14 Aug 2007, giuseppe.derossi@email.it wrote:
>
> > Name, Surname , Job, Hobby, Pet, address
> > John, Smith,   NULL, photo, NULL, NULL
> > John, Smith,   student, NULL, cat, NULL
>
> > by using name and surname as selecting key, I want :
> >
> > John, Smith,   student, photo, cat, NULL
>
> If you are sure there's no conflict (multiple values) for a (name,
>    surname) key, one simple solution might be (surely not an efficient
one)
>
> select
> (select distinct Name from thetable where Name = 'name1' and Surname =
'surname1' where Name is not null),
> (select distinct Surname from thetable where Name = 'name1' and Surname =
'surname1' where Surname is not null),
> (select distinct Job from thetable where Name = 'name1' and Surname =
'surname1' where Job is not null),
> (select distinct Hobby from thetable where Name = 'name1' and Surname =
'surname1' where Hobby is not null),
> (select distinct Pet from thetable where Name = 'name1' and Surname =
'surname1' where Pet is not null),
> (select distinct address from thetable where Name = 'name1' and Surname =
'surname1' where address is not null)
>
> This will fail if you also have
> John, Smith,   student, NULL, dog, NULL
>
> in addition to
> > John, Smith,   student, NULL, cat, NULL
>
> You can test for offending rows by:
>
> select
> array_to_string( array( select distinct Pet from thetable where Name =
'name1' and Surname = 'surname1' where Pet is not null),',')
>
> and so on...
>
>
> Regards,
>
> Ben K.
> Developer
> http://benix.tamu.edu
>
> ---------------------------(end of broadcast)---------------------------
> TIP 5: don't forget to increase your free space map settings
>
>
>
 --
 Email.it, the professional e-mail, gratis per te: http://www.email.it/f

 Sponsor:
 Hai bisogno di contanti per realizzare i tuoi desideri? Prometeo ti propone
prestiti da 1.500 a 31.000 Euro! Clicca qui per un preventivo immediato.
 Clicca qui: http://adv.email.it/cgi-bin/foclick.cgi?mid=6916&d=20070814

>>> On Tue, Aug 14, 2007 at  8:48 AM, in message
<b1a365afcf3059983eecc3db74a3cadf@83.225.75.94>, <giuseppe.derossi@email.it>
wrote:
> I was studying a solution which foresees the 'case when' construct...
> now I can use it in addition to  array_to_string . In fact my solution
> failed too if <it's raining cats and dogs :-) >, but if there are some more
> values I get them all as well.

Is there a timestamp, date, or sequence number that you can use to break
ties?  Without something like that, I don't think you can do what you want
with set-based logic -- you would need to do something procedural without
any guarantee that you'd get the same results in two different runs.

-Kevin