Problem with complex select/IDS 7.31.UC3 & statistics
Hi, I'm just playing around with a testing environment because we're facing a weird problem. We've got IDS 7.31.UC3 running on a Solaris 2.6 box; after issuing a complex select statement with several tables (some of them outer joined) we're getting a result set of 73 rows though we were expecting only one row (which is correct, because the 73 rows IDS provides are completely identical!)! This doesn't happen on 7.23.TC14 (NT4/SP4) and 7.24.UC5 (Solaris 2.6). Now, I've set up a small database only with the tables used in the select and loaded the data to achieve the correct result set. The first try succeeded and the select delivered one row! After wondering why it worked this time I set up the statistics (low w/ drop distributions) and I retrieved 73 rows! Now, maybe somebody out there encountered the same problem and, btw, how can I get rid of the statistics completely for testing purposes because dropping and creating the database over and over again doesn't make me happy... ;-) TIA and many regards, Stephan. -- Stephan Stresing MR Informatik GmbH mailto:st@mr-informatik.de Sent via Deja.com http://www.deja.com/ Before you buy.
In article <7v1rh1$79i$1@nnrp1.deja.com>, Stephan Stresing <stephan_stresing@my-deja.com> writes >Hi, >I'm just playing around with a testing environment because we're facing >a weird problem. We've got IDS 7.31.UC3 running on a Solaris 2.6 Is this 7.31.UC3-1 ?? >box; after issuing a complex select statement with several tables (some >of them outer joined) we're getting a result set of 73 rows though we >were expecting only one row (which is correct, because the 73 rows IDS >provides are completely identical!)! This doesn't happen on 7.23.TC14 >(NT4/SP4) and 7.24.UC5 (Solaris 2.6). Now, I've set up a small database >only with the tables used in the select and loaded the data to achieve >the correct result set. The first try succeeded and the select delivered >one row! >After wondering why it worked this time I set up the statistics (low >w/ drop distributions) and I retrieved 73 rows! Now, maybe somebody out >there encountered the same problem and, btw, how can I get rid of the >statistics completely for testing purposes because dropping and creating >the database over and over again doesn't make me happy... ;-) >TIA and >many regards, > Stephan. > >-- >Stephan Stresing > >MR Informatik GmbH >mailto:st@mr-informatik.de > > >Sent via Deja.com http://www.deja.com/ >Before you buy. -- David Williams
In article <QFV6uNA+2NF4Ewih@smooth1.demon.co.uk>, David Williams <djw@smooth1.demon.co.uk> wrote: > In article <7v1rh1$79i$1@nnrp1.deja.com>, Stephan Stresing > <stephan_stresing@my-deja.com> writes > >Hi, > >I'm just playing around with a testing environment because we're facing > >a weird problem. We've got IDS 7.31.UC3 running on a Solaris 2.6 > Is this 7.31.UC3-1 ?? Yes, it is (according to the sticker on the package). Is this a sign for an exploding server within a short period of time??? ;-))) TIA, Stephan. > > >box; after issuing a complex select statement with several tables (some > >of them outer joined) we're getting a result set of 73 rows though we > >were expecting only one row (which is correct, because the 73 rows IDS > >provides are completely identical!)! This doesn't happen on 7.23.TC14 > >(NT4/SP4) and 7.24.UC5 (Solaris 2.6). Now, I've set up a small database > >only with the tables used in the select and loaded the data to achieve > >the correct result set. The first try succeeded and the select delivered > >one row! > >After wondering why it worked this time I set up the statistics (low > >w/ drop distributions) and I retrieved 73 rows! Now, maybe somebody out > >there encountered the same problem and, btw, how can I get rid of the > >statistics completely for testing purposes because dropping and creating > >the database over and over again doesn't make me happy... ;-) > >TIA and > >many regards, > > Stephan. > > > >-- > >Stephan Stresing > > > >MR Informatik GmbH > >mailto:st@mr-informatik.de > > > > > >Sent via Deja.com http://www.deja.com/ > >Before you buy. > > -- > David Williams > -- Stephan Stresing MR Informatik GmbH mailto:st@mr-informatik.de Sent via Deja.com http://www.deja.com/ Before you buy.
Hi again, any new ideas concerning my problem? Our Informix support team is still investigating, too, and I hoped to get some more input from you in order to speed up the solution process... TIA and many regards, Stephan. P.S. Hi David, why did you ask for 7.31.UC3-1? My comment wasn't meant to offend anybody, just my (insane) kind of humour... :-)) I'm really interested in your reasons for this specific question - is there anything wrong with this version of IDS? In article <7v6egq$i1g$1@nnrp1.deja.com>, Stephan Stresing <stephan_stresing@my-deja.com> wrote: > In article <QFV6uNA+2NF4Ewih@smooth1.demon.co.uk>, > David Williams <djw@smooth1.demon.co.uk> wrote: > > In article <7v1rh1$79i$1@nnrp1.deja.com>, Stephan Stresing > > <stephan_stresing@my-deja.com> writes > > >Hi, > > >I'm just playing around with a testing environment because we're > facing > > >a weird problem. We've got IDS 7.31.UC3 running on a Solaris 2.6 > > Is this 7.31.UC3-1 ?? > > Yes, it is (according to the sticker on the package). Is this a sign for > an exploding server within a short period of time??? ;-))) > TIA, > Stephan. > > > > > >box; after issuing a complex select statement with several tables > (some > > >of them outer joined) we're getting a result set of 73 rows though we > > >were expecting only one row (which is correct, because the 73 rows > IDS > > >provides are completely identical!)! This doesn't happen on 7.23.TC14 > > >(NT4/SP4) and 7.24.UC5 (Solaris 2.6). Now, I've set up a small > database > > >only with the tables used in the select and loaded the data to > achieve > > >the correct result set. The first try succeeded and the select > delivered > > >one row! > > >After wondering why it worked this time I set up the statistics (low > > >w/ drop distributions) and I retrieved 73 rows! Now, maybe somebody > out > > >there encountered the same problem and, btw, how can I get rid of the > > >statistics completely for testing purposes because dropping and > creating > > >the database over and over again doesn't make me happy... ;-) > > >TIA and > > >many regards, > > > Stephan. > > > > > >-- > > >Stephan Stresing > > > > > >MR Informatik GmbH > > >mailto:st@mr-informatik.de > > > > > > > > >Sent via Deja.com http://www.deja.com/ > > >Before you buy. > > > > -- > > David Williams > > > > -- > Stephan Stresing > > MR Informatik GmbH > mailto:st@mr-informatik.de > > Sent via Deja.com http://www.deja.com/ > Before you buy. > -- Stephan Stresing MR Informatik GmbH mailto:st@mr-informatik.de Sent via Deja.com http://www.deja.com/ Before you buy.