paul-rogers opened a new pull request, #2867: URL: https://github.com/apache/drill/pull/2867
# Support for non-projected complex vectors ## Description The EVF mechanism provides scan-time projection for many vector types. The reader code is simple: it deserializes all columns for formats such as JSON, CSV, etc., and writes them to the `ColumnWriter` objects. Internally, EVF simply ignores the data for unprojected columns. This solution simplifies the readers: it is not necessary for each reader to include the complex code to handle projection. This solution is also performant: projection is done at scan time rather than the other approach, which is to read all data into vectors, then allow a PROJECT operator to drop the unprojected columns. Present EVF projection support handles most scalar and "well-structured" columns (repeated types AKA arrays, maps, etc.) However it does not handle the more esoteric types UNION, LIST (AKA repeated UNION), REPEATED LIST (AKA repeated, repeated UNION). This PR provides more support, though holes remain. ## Documentation This is an internal feature: no user-visible documentation is required. ## Testing Extended existing EVF-related unit tests. -- 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. To unsubscribe, e-mail: dev-unsubscr...@drill.apache.org For queries about this service, please contact Infrastructure at: us...@infra.apache.org
