You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
What? — It dumps the existing (0.14) routines (ops/functions/aggregations) to a definitions file a’la substrait extensions.
Why? — We have to redefine all of the functions on PType now, so this is our indirection (which partiql-lang-rust may leverage)
This commit writes the current function signatures and variants to Ion files (like substrait extensions) so that PartiQL function signatures can be expressed independently of implementations. I am using these for my PR which moves functions out of SPI and into planner/eval respectively.
Open Questions
What might be some limitations of this representation?
Are the types sane for defining inputs/outputs?
How might packaging/namespacing fit into this? i.e. why even have sql99::… and partiql::… namespaces?
Is there value in these being public and consumed across PartiQL implementations, or is it just a nice-to-have for plk developers.
Where might these live long term if shared?
How are signatures stringified (specific/key) for representing in a plan?
License Information
By submitting this pull request, I confirm that my contribution is made under the terms of the Apache 2.0 license.
The text was updated successfully, but these errors were encountered:
Relevant Issues
Description
This commit writes the current function signatures and variants to Ion files (like substrait extensions) so that PartiQL function signatures can be expressed independently of implementations. I am using these for my PR which moves functions out of SPI and into planner/eval respectively.
Open Questions
License Information
By submitting this pull request, I confirm that my contribution is made under the terms of the Apache 2.0 license.
The text was updated successfully, but these errors were encountered: