tokoko commented on code in PR #12800:
URL: https://github.com/apache/datafusion/pull/12800#discussion_r1792879962
##########
datafusion/substrait/src/logical_plan/consumer.rs:
##########
@@ -781,10 +757,11 @@ pub async fn from_substrait_rel(
from_substrait_named_struct(named_struct, extensions)?
.replace_qualifier(table_reference.clone());
+ let substrait_schema =
+ apply_projection(substrait_schema, &read.projection)?;
Review Comment:
that's a good point. TableScan in DF seems to have a `filters` field, which
I'm assuming is supposed to work the same way as substrait's filter, so I don't
think we're in danger of doing that. Although this probably means that even the
part of the base_schema that is being masked out needs some sort of validation
because ReadRel's filter (and/or best_effort_filter) field might operate on
them. This probably decides the question I raised above whether plans with
inconsistent data types in masked out fields should be accepted or not.
--
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]
