Travis Athougies
Travis Athougies
Indeed you are correct. Will investigate a solution.
I think this is a good idea. Someone at one point also was interested in a 'binary' PG backend, and it would be good to share code between the two.
What does it do now?
binary in `beam-migrate` refers to the standard `BINARY` type. I think these should map to `ByteString`, not `Integer`. And then, you're saying we can achieve this behavior on SQLite using...
This is _somewhat_ intended behavior due to the difficulty of dealing with nullables in SQL... We can't determine if the Maybe came from the fact the row didn't exist, versus...
Perhaps we could move it to a separate 'utility' module, and then make it return a more obvious error. I like the `Either MoreRowsError row` idea. Thoughts?
Hello there... sorry for the long delay. I think this is a good change, but I worry that `autoMigrate` will now not fail on an unsafe migration.
Hey everyone, There is currently no way to do this, but I'm happy to accept PRs. You'd need to add a function that could take a composite field name and...
Do we have a list of where it's ignored? It should be used?
The current best way to work around this is to use `unsafeRetype_`. The behavior you're seeing is incredibly annoying but apparently correct and documented in postgres: https://www.postgresql.org/docs/13/functions-aggregate.html (Apparently postgres thinks...