You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
However (due to constraints and transformations), the column may no longer exist once we hit the fitting stage -- meaning that my specified numerical distribution won't be applied. For example, if I apply an Inequality constraint between start_date and end_date, then start_date may be dropped.
Expected behavior
It would be nice to log that this is happening (logger.info):
After preprocess, we should have a new list of column names. If any user-provided name in numerical_distributions is no longer available, then log it during the fit_processed_data method:
INFO: Requested distribution ('beta') cannot be applied to column ('start_date')
because it no longer exists after preprocessing.
The text was updated successfully, but these errors were encountered:
Nice to have. We can remove this from the milestone if needed.
Problem Description
In SDV 1.0, I will be able to specify numerical distributions for any original column name:
However (due to constraints and transformations), the column may no longer exist once we hit the fitting stage -- meaning that my specified numerical distribution won't be applied. For example, if I apply an Inequality constraint between
start_date
andend_date
, thenstart_date
may be dropped.Expected behavior
It would be nice to log that this is happening (
logger.info
):After
preprocess
, we should have a new list of column names. If any user-provided name innumerical_distributions
is no longer available, then log it during thefit_processed_data
method:The text was updated successfully, but these errors were encountered: