embucket-labs
embucket-labs copied to clipboard
Embucket
Experimental version. A BYOC option for Snowflake workloads
Support `...::BINARY` & `CAST(... AS BINARY)` via a `to_binary` UDF call
Support `...::CHAR` & `CAST(... AS CHAR)` via a `to_char` UDF call
Support `...::BOOLEAN` & `CAST(... AS BOOLEAN)` via a `to_boolean` UDF call
Support `...::NUMBER(precision, scale)` & `CAST(... AS NUMBER(precision, scale))` via a `to_time` UDF call
Support `...::TIME` & `CAST(... AS TIME)` via a `to_time` UDF call
Support date input format for `to_date` UDF and cast via (or): - Akin to `to_timestamp` impl - Creating an Arc map for context vars - Some other way (DF `SessionContext`...
I use this SQLAlchemy URL to connect Superset to Embucket: `snowflake://embucket:XXXXXXXXXX@test/data_platform_db?host=embucket&port=3000&protocol=http&schema=data_platform_schema&warehouse=data_platform_db` (password is hidden by XXX...) It clearly sets the database as `data_platform_db`. This database is created with an s3...
### This is how it works with Snowflake https://github.com/user-attachments/assets/2a68fdf0-01df-43aa-b087-d9c4c43190de ### This is how it works with Embucket https://github.com/user-attachments/assets/2a04b944-008f-43d0-b6cf-16b379b24e02 This is the log from Embucket during this error: ``` {"timestamp":"2025-07-25T14:17:34.053980Z","level":"INFO","fields":{"message":"Creating DedicatedExecutor"},"target":"core_executor::dedicated_executor"}...
This is the Epic for all the issues related to Embucket usage in Superset.
## Error While working on `to_date` and `try_to_date` UDFs (PR: #1376; Issues: #624 & #1156). I got this error multiple times (similar ones), both in SLT tests and insta tests:...