Hi All,

We have a project requirement to upload around 54 million records (available in the form of a csv file) into a DB2 table.
The destination table is empty, has 37 columns (all simple data types – no LOBs) and has no column constraints or composite keys.

We tried this activity with the import/load command, but it took us around 6 hrs to load 4 million records! DB2 Data studio did not help us as well.

Can someone please let us know of any easier/faster method to bulk load such a large volume of data in DB2.

Thanks
Chandra.

buy more iron.

Mate,

u got to be kidding me to have posted such a humorous reply..

jokes apart..

can u pls provide me with some really technical solution, as we need to find a way to get this Bulk load thing completed at the earliest?

Cheers
Chandra.

what is the 'purest' input format for data to the load utility?

a csv file will work,
but does the utility have to perform some kind of 'reformatting'
before the actual loading to the tables takes place?

apparently this is a 'one-off' exercise.
either bite the bullet and do the load as is
or
'normalize' your csv file to a type that requires no formatting by the utility - which will also take time - but may make the load process run faster.

Are you sure you are using LOAD? I have never seen it take that long to load that few records. Post your command.

Look at the disk-layout and tablespaces. No SMS I hope? No raid5 I hope? You need the best write-performance and DMS tablespaces already in the right size and spread over as many disks as possible. And "util_heap_sz " max.size?