Cognos Analytics

 View Only

 Using DB2 vs Parquet for Data Sets

Michael Bernier's profile image
Michael Bernier posted Thu February 26, 2026 01:57 PM

Hello,

We're fairly new to using Cognos (started with 12.0.0 last year), but discovered early on the value of using data sets to improve overall performance compared to data modules for our larger business users. While we've been using the default Parquet files with good results, I'm curious if there are any advantages to using a DB2 database for hosting data sets (aside from eliminating the 2Gb size limit for Parquet files)? We're primarily a Microsoft SQL shop so our experience with DB2 is mostly nonexistent, so go easy on me 🙂

Jorge Carreon's profile image
Jorge Carreon

Hello,

Moving datasets to DB2 or any other database other than default Parquet files, will help you releasing memory that was been used by those files. We have come across the same issue where parquet files started consuming a lot of memory from our Cloud instance, and after months of health checks and support reviews, it all pointed to move away from parquet files specially if you have large data, and have them store in a data warehouse to be accessed as you mentioned.

I hope that helps, please let me know if I can help in anyway.

Best Regards

Jorge C

Patrick Neveu's profile image
Patrick Neveu IBM Champion

Hi Michael,

The benefits of using Db2 include:

  • Reduced workload on Content Manager, which doesn’t receive, return, or store Parquet files.
  • The query execution is offloaded from the Cognos Analytics application tier to the Db2 server.

    The Compute service is used only when Parquet files are required.

  • The Content Manager size restrictions are not applicable.

    Data is stored in a separate database instance and dependent on the storage space and compute capacity of the Db2 server.

You will find a complete documentation in here: Storing data from uploaded files and data sets in Db2 - IBM Documentation

Best regards,

Patrick Neveu