clintropolis commented on issue #14339:
URL: https://github.com/apache/druid/issues/14339#issuecomment-1707295891
>Thanks @clintropolis! Maybe I'm missing something, but how to set auto in
this ingestion query:
Ah, this is the MSQ query context thing I was referring to, adding
`{"useAutoColumnSchemas": true}` to the query context is required in this case,
you leave the type definition as `COMPLEX<json>`. 'auto' and 'json' can both
produce the `COMPLEX<json>` logical type, however internally they use different
physical formats in the segment files, and by default `COMPLEX<json>` still
maps to the older 'json' version so the flag is necessary to get the new
version.
One thing to be aware of, when using `useAutoColumnSchemas`, all built-in
types will be written with the 'auto' column writers, which will make some
differences when the flag is not set, the main difference is that numeric types
like `LONG`/`BIGINT` will be written out with value indexes, which will make
the segments larger, but enable faster filtering. Currently the flag is 'all or
nothing', but in the future we will probably introduce some additional control
over this in one way or another, still working out some of the details.
--
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: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
