2023-10-19 kernel meeting notes #55
zachschuermann
started this conversation in
General
Replies: 0 comments
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Uh oh!
There was an error while loading. Please reload this page.
Uh oh!
There was an error while loading. Please reload this page.
-
summary
Lightweight sync then chatted about #26
action items:
attendees
@rtyler @wjones127 @nicklan @MrPowers @zachschuermann @vkorukanti @roeap @ryan-johnson-databricks
notes stream
in-flight PRs
from @roeap: dagster in prod!
nullability handled in types themselves vs systems that use the types. nullability should be expressed as
Expression::Null
and that's it. should we have 'typed' null? e.g. the difference between u32 Null and u64 Null.expression evaluation: should probably enforce expected return schema? only take input schema https://github.com/delta-incubator/delta-kernel-rs/blob/main/kernel/src/lib.rs#L112
if we take input/output schema: how disruptive would that be? if we get to a point where we (kernel) cannot determine output schema that's a red flag. for now: col pruning, stats/file skipping, DVs, row_id generation, constraints, column mapping, actual query schema we handoff: engine should tell us the schema. all prior examples have predetermined output schema. from @wjones127 we could have API be
evaluate
andget_output_schema
. from @ryan-johnson-databricks we can do both (kernel/connector schema validation).expressions
Beta Was this translation helpful? Give feedback.
All reactions