jack86596 commented on pull request #3924: URL: https://github.com/apache/carbondata/pull/3924#issuecomment-706907380
The main idea of this PR is to improve the performance of following query: select * from table where col1 = 'xxx' or col2 = 'yyy'; where sort_columns order is col1,...,col2. Currently we can only create SI on col2, but if only 1 column SI is created another is not, above query will still just scan the main table directly, will not use col2 SI, leads to bad performance. So we have two options to improve this kind of query: 1. create SI on col1 also, this is the solution of this PR 2. if only one SI is created, still use this SI, other columns use main index to prune, this will be done later. And for no_sort table, first column is not sorted either, main index is worse than SI, under this scenario, we need this PR to create SI on the first column. ---------------------------------------------------------------- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org