>
>Hi Benjamin, Arjen
>
>I did some testing again. I have 75,000,000 rows table, after using group 
>by I got 62,300,000 rows.
>If I use *nix -u sort, I got 74800,00 rows.
>
>I compared the two results, I feel I trust my *nix result. Some lines just 
>miss in "group by' result, look like no reason at all.
>I think this is a bug.
>If you like to see my two result, I can attach to you.
>
>Thanks.
>
> >Arjen,
>
> >I don't have any NULL field in my table, but I have empty field which is 
> ' ' in my some fields.
> >Is that reason the different between SELECT DISTINCT and *nix sort?
>
> >If the answer is yes, I have to dump the table out then use the *nix sort.
>
> >Thanks
>
>At 01:05 PM 7/24/2002 +1000, you wrote:
>Hi Benjamin, Vivian,
>
>On Wed, 2002-07-24 at 05:23, Benjamin Pflugmann wrote:
> > On Tue 2002-07-23 at 14:58:16 -0400, [EMAIL PROTECTED] wrote:
> > > I have 14 fields are chars (like char(8), char(9)), and one field is 
> int(4).
> > > All char fields are mixed with upper case letters, numbers(0-9) and
> > > sign(like $).
> > > Because there is no low case char, I don't have to put -f option for 
> unix
> > > sort.
> >
> > Well, it never hurts to try anyhow... *hint*
> >
> > Does the query use indexes? If so, please check they are not corrupted
> > (e.g. using CHECK TABLE).
> >
> > Which character set is MySQL set to, which "locale" is set in your
> > shell? Maybe the collections are different and consider different
> > characters the same.
> >
> > Please try to find some rows which "disappear". For example, dump the
> > result from the query (62,000,000 rows) to a file like you have done
> > for the 76,500,000 rows and find out which rows got removed
> > (e.g. sorting the files and using "comm" on the 76,000,000 and
> > 62,000,000 version).
> >
> > In short, try to make a reproducable test case. Thanks.
>
>Note that SELECT DISTINCT only grabs rows where none of the distinct
>columns is NULL.
>
>A *nix sort on a dump would not do the same. Try doing a "grep -v NULL"
>first before sorting. Then see what you have left.
>However, if your table dump also contains NULLs in fields that are not
>part of the DISINCT selection, then you'll need to make the filtering
>slightly more complicated to have it produce an equivalent result.
>
>
>Regards,
>Arjen.
  


---------------------------------------------------------------------
Before posting, please check:
   http://www.mysql.com/manual.php   (the manual)
   http://lists.mysql.com/           (the list archive)

To request this thread, e-mail <[EMAIL PROTECTED]>
To unsubscribe, e-mail <[EMAIL PROTECTED]>
Trouble unsubscribing? Try: http://lists.mysql.com/php/unsubscribe.php

Reply via email to