Strange performance of IN-Clause Select

[moeschl]

Hi
I have experienced a strange behaviour of my DB2 DB when executing a IN-Clause Select on a very well indexed table.
With 35 objects in the IN-Clause it takes almost one minute to show results, with 34 or 36 objects it finishes after millisecs.
What can be the reason for that strange effect ?
greetings from bavaria's oktoberfest !

 

[stoggers]

Hi,

Try using DB2 explain to determine what access paths are being used...

Regards,

Stoggers.

 

[moeschl]

Hi

It uses exactly the same access-path ! (34,35 or 36 objects)
can there be a kind of internal splitting of the statement or a chunk-value like in oracle ?

 

[MarcLodge]

Are the IN clauses all the same size? What is the total length of the 34/35 and 36 clauses?

 

[moeschl]

Hi
Problem solved automatic after patching.
Seems to have been a "special feature".
Access-pathes have been equally for 34,35 or 36 elements, but only 35 were performing bad. Every element was a 6 byte literal. nothing in the patches description referred to be solving a problem like this, but its gone...thanks for your engagement, anyway.