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
{{ message }}
Repository navigation
Feature Request: Enhanced Soft Failure support #122
Just started experimenting with dataframely and liking a lot of what I see.
It would be great if Soft Failure support was extended in a way that makes it easy to generate failure reports for users who supply the input data.
Some ideas that sound useful to me:
Provide a way to mark a set of columns as debug ID columns for Failure reports
Extend FailureInfo counts and cooccurrence_counts methods with a "group_by_ids=True" or "group_by='id'" or similar parameter - if set, errors are grouped by the id columns suggested above
it would be great if friendly names (that have a meaning to non technical users) could assigned to dy.rule
raayu83 could you elaborate what's currently missing to assign "friendly names" to rules? The rule assumes the name of the function which allows you to provide arbitrarily expressive names.
What I mean is an alternative representation of the Name that looks cleaner and allows you to use any character you want.
I'd like to add that I see this as a nice bonus feature, nothing really important to get things done.
Fair enough. I am personally not a fan of adding this as it will allow providing arbitrary names to @dy.rules but you're still restricted to auto-generated names for column-level checks (e.g. nullability, min, max, ...). Maybe we can reconsider this at some point as dataframely's usage area expands but I will close this for now.
Activity
borchero commentedon Feb 17, 2026
raayu83 could you elaborate what's currently missing to assign "friendly names" to rules? The rule assumes the name of the function which allows you to provide arbitrarily expressive names.
raayu83 commentedon Feb 18, 2026
Hi Oliver Borchert (@borchero) ,
What I mean is an alternative representation of the Name that looks cleaner and allows you to use any character you want.
I'd like to add that I see this as a nice bonus feature, nothing really important to get things done.
borchero commentedon Feb 18, 2026
Fair enough. I am personally not a fan of adding this as it will allow providing arbitrary names to
@dy.rules but you're still restricted to auto-generated names for column-level checks (e.g. nullability, min, max, ...). Maybe we can reconsider this at some point as dataframely's usage area expands but I will close this for now.borchero commentedon Feb 18, 2026
Ah prematurely closed this, I only meant to close the "friendly names" part. Will keep open.