adriangb commented on code in PR #19137:
URL: https://github.com/apache/datafusion/pull/19137#discussion_r2598687754
##########
datafusion/datasource/src/table_schema.rs:
##########
@@ -70,7 +70,7 @@ pub struct TableSchema {
///
/// These columns are NOT present in the data files but are appended to
each
/// row during query execution based on the file's location.
- table_partition_cols: Vec<FieldRef>,
+ table_partition_cols: Arc<Vec<FieldRef>>,
Review Comment:
I tried this but it:
1. Is incompatible with your recommendation in
https://github.com/apache/datafusion/pull/19137/files/fd79a3ba89e9c5d45796aa8fbb7782bf1413bde4#r2598642330
because we can't do `Arc::make_mut(&mut Fields)` and `Fields` doesn't expose
mutablility.
2. It would require changing the public API of
`TableSchema::partition_fields()`
I also am not that worried about construction cost (that happens once in
`FileSource`, it's the mutliple copies for each `FIleOpener` that worry me.
In any case since the current change is fully backwards compatible API wise
and is internally contained we can always come back and use `Fields` later, so
I will leave that as a followup.
--
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]
