Closed nikhil96sher closed 1 year ago
@mIXs222 Can you review this and other wanderjoin query related PR(s).
Also i'll merge those wanderjoin PRs together then. Thanks for taking over them!
Nvm, seems like you could initiate the merges from this PR
The cardinality of lineitem table as 6_000_000 * SF is approximate. For some scales, the actual cardinality is smaller than this value (by a very small number). Owing to this, the cardinality ratio in the last block is not 1.0 and hence the final result is still approximate (since scaling is performed). Currently, fixed it by setting the cardinality of the last dataframe read to be 1.0.
Let me know if you have other suggestions!