We Want to Hear From You!

Would like to know how to reduce the disk space - using encoding , partitioning etc in detail.

It'd be nice to see some insights into executing the DBD process, tips/tricks, etc. We identified a problem with exporting the projection design via DBD in management console, Support said a JIRA Bug was created (VER-58419) but I haven't seen or heard if it's been fixed. Because of that we're executing through AdminTools which does not have the same features as MC version (e.g., no option for correlated columns).

@Vertica_User - Encoding can help reduce disk usage, but partitioning won't

See:
https://my.vertica.com/docs/9.0.x/HTML/index.htm#Authoring/AnalyzingData/Optimizations/OptimizingEncodingToImproveQueryPerformance.htm

https://my.vertica.com/kb/Redesigning-Projections-for-Query-Optimization/Content/BestPractices/Redesigning-Projections-for-Query-Optimization.htm

@ScottL - I did a test in MC 9.0.1.

The "Export Design" button is enabled if the Design is built but not deployed.

The "Export Design" button is disabled if the Design is deployed.

I attached images that show this.

Also, you can manually run the ANALYZE_CORRELATIONS function to analyze the specified tables for pairs of columns that are strongly correlated.

See:
https://my.vertica.com/docs/9.0.x/HTML/index.htm#Authoring/SQLReferenceManual/Functions/VerticaFunctions/ANALYZE_CORRELATIONS.htm

Thanks!

Thanks, Jim - they may have fixed it in 9; we're running 8.1 and will be upgrading to 9 in the next few months. Still, if they're looking for blog material, it would be great if someone could share insights/tips/tricks into executing the DBD process. Thanks for your help!

Thanks for these suggestions, we are working on content and will notify you when it is available! Much appreciated!

Hi!

If there is a topic you would like to see in an upcoming blog post or best practices document, let us know in the comments.

Would be nice to read - "UDF best practice". Resources allocation, how to analyze resource usage(memmory, cpu, network) for UDF, best practice for parallelization, etc.