Devin Smith
Devin Smith
Or maybe, makes it the responsibility of the JS layer; still, not sure if we'd want to unconditionally unwrap parent of said table.
A test skeleton has been sketched out for this in `io.deephaven.iceberg.PyIceberg1Test#cities2`
Our `io.deephaven.parquet.base.RowGroupReader#getColumnChunk(java.lang.String, java.util.List)` smells; we should not need to do the resolution mapping per column per row group; it should only need to happen once per column. Ie, a given...
https://github.com/apache/iceberg/blob/main/format/spec.md#schema-evolution should be consulted in general for issues about how schema evolution and tracking is supposed to happen. https://www.mail-archive.com/[email protected]/msg02236.html also talks a bit about Iceberg field-id and Parquet field_id.
Potentially related to #6124
This may be a separate issue? I'm not sure if we need to support proxied context roots?
A temporary workaround might be to translate the LocalDate into an Instant, either on the server, or from the client before subscribing. ```python my_table = ( my_table .update_view( [ "my_timestamp=toInstant(my_date,...
@dependabot rebase
@dependabot rebase
@dependabot rebase