RE: is it a bug? - Mailing list pgsql-general

From Francis Solomon
Subject RE: is it a bug?
Date
Msg-id NEBBIFFPELJMCJAODNPKKEFECDAA.francis@stellison.co.uk
Whole thread Raw
In response to is it a bug?  ("He weiping" <laser@zhengmai.com.cn>)
List pgsql-general
Hi,

I think you've misunderstood the purpose of the functions.
They exist to *pad* the strings, not to truncate them.
Your examples will both return 'laser' because char_length('laser') = 5
and you asked for a padded length of 4.

Had you done this: select lpad('laser', 8, '*');
You would get this: ***laser

... and obviously with rpad() you would have seen 'laser***' instead.

If you want to truncate strings, try this:
select substring('laser' from 1 for 4);
... which will truncate to length 4, i.e. 'lase'

I couldn't find a combination function that will perform both of these
functions in one. However, you could try a construct like this:

select rpad(substring('laser' from 1 for xx), xx, '*');

... where 'xx' is the number of characters you want in the final string.
I'm sure you could wrap a user-defined function around this to that
you'd only have to feed in the number of characters once instead of
twice. Perhaps someone else knows a better way of doing this?

Hope this helps

Francis Solomon

>I'm using cvs-current, and testing those build-in function
>according to the docs.
>but it seems the "lpad", "rpad" don't work,
>when I type:
>select lpad('laser', 4, 'a');
>in psql, the result is still
>'laser', the same with 'rpad',
>Is it a bug or I'm mis-understaning the lpad and/or rpad functions?
>
>Regards
>
>Laser


pgsql-general by date:

Previous
From: Anand Raman
Date:
Subject: insertion times ..
Next
From: "Martin A. Marques"
Date:
Subject: Postgres BLOBS and PHP