If a modelling project has a senscase list of e.g.:
1
2
3
HIGH
MID
LOW
4
5
6
this will fail on .parquet saving with pyarrow due to mixed types in same column.
Best solution would probably be to use a priori knowledge wrt. columns that should always be considered strings (e.g. SENSCASE), and instruct the data input layer to always have these columns as string type.
If a modelling project has a senscase list of e.g.:
this will fail on
.parquet
saving withpyarrow
due to mixed types in same column.Best solution would probably be to use a priori knowledge wrt. columns that should always be considered strings (e.g.
SENSCASE
), and instruct the data input layer to always have these columns as string type.