Hi all,

I'm new to DB2 and I my proc has to process some 3 million records. Can I use cursors for this. Will cursors handle such large volumes of data efficiently or will it hang coz it need to store all 3 million rec in temp memory. Do we have any other alternative in DB2 other than using cursors to process such huge data? Pls advise.


Thanks,
Kayal.

all rows will not be loaded into memory
result sets will be created in temp. tablespace
I don't see how to handle this in other way..
it will take some time if the rows are all handled one by one ..

Hi
Thanks for the reply. In my cursor I dont have any condition to get the result set, so entire table has to be handled by the cursor. Is there any spl keywords in cursors to fetch the result set every one lakh record and process and again fetch 1 lakh into memory?

Thanks,
Kayal

if there is no predicate it will fetch from base table i presume and lock according the isolation level...
db2 will always bufferpools to transit the data
have you already run this process ?
is there a specific problem ?
be as precise as possible and add some additional info if so..

What do you want to do with those 3 mio rows? Maybe you can push the processing into the database system and just retrieve the final result (or something close to it).

However, if you want to do the processing in your application yourself, this has a high chance to be simply slow.