UDTFs: Default column names updated for Python vectorized UDTFs¶
This behavior change is in the 2023_06 bundle.
For the current status of the bundle, refer to Bundle History.
This change only affects users of Vectorized Python UDTFs (user-defined table functions).
The default column names for the input DataFrame to a vectorized UDTF are the indices, such as 0,1,2, … etc.
The default column names for the input DataFrame to a vectorized UDTF match the signature of the SQL function. Using numeric indices by default will no longer work and you will get a “Key Error”.
The column names will follow the SQL identifier requirements. Namely, if an identifier is unquoted it will be capitalized, and if it’s double quoted it will be preserved as it is.
For details, see Vectorized Python UDTFs.