-
Notifications
You must be signed in to change notification settings - Fork 8
Open
Description
Hi,
I develop a Haskell project Napkin, which talks to multiple database backends (Postgres, Sqlite, BigQuery, SqlServer). There is an issue with having unified API of SQL functions. Every backend has lots of differences. Sqlite is very simple just a few dozen functions.
Could you help me with estimation how well Hoogle+ might reconstruct PostgreSQL functions based on a set of Sqlite functions?
PostgreSQL API is formally described through meta table (function name, immutability, argument types etc).
List of Sqlite functions is not long and might be curated manually. So ML approach looks good at first glance, doesn't it?
Metadata
Metadata
Assignees
Labels
No labels