Replies: 2 comments 1 reply
-
Bumping this, we're working with bytes in a column in python, and when I use to_df() I get a pd.Series column with dtype=np.object instead of dtype=bytes on the other side. This means I have to do some extra work to convert it back to a dtype=bytes. |
Beta Was this translation helpful? Give feedback.
1 reply
-
This would be great. I would like to go the other way too (from 16-byte blob to UUID). |
Beta Was this translation helpful? Give feedback.
0 replies
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
-
Casting UUID to blob is currently not supported. We need this to align between parquet files (with row-level UUIDs) generated by duckdb and parquet files generated outside duckdb. We want duckdb to return a
Buffer
in node.js (instead ofstring
) when reading the UUID column, regardless if the column was generated using manual python or sql viauuid()
.To solve this we need to cast uuid to blob:
select uuid()::BLOB;
We currently get:
Error: Conversion Error: Unimplemented type for cast (UUID -> BLOB)
Casting should hopefully be trivial since UUID is backed by a 16 byte blob.
Beta Was this translation helpful? Give feedback.
All reactions