Closed Burhanuddin-Nahargarwala closed 4 months ago
Hi @Burhanuddin-Nahargarwala thanks for filing the issue.
Seems like this is actually an issue in our documentation for this constraint. Please use the keyword 'strict_boundaries'
, not 'strict'
.
Note that the parameters are listed correctly at the top of the docs page -- it is just the example that's incorrect. We will update the example in the docs.
To re-surface our message from Slack: By default, SDV synthesizers will enforce that all the synthetic data will adhere to the same min/max boundaries as observed in the real data. This means that as long as your real data has values >=0, then the synthetic data will as well — no constraint is needed
I realize you are adding the constraint just in case there end up being negative values (because you want to know when this happens). If at all possible, I would recommend performing such a check yourself before using SDV. Constraints are only meant to be used if you don't have any other options available -- as adding them may impact performance and quality in different ways.
Docs are now updated. (You may need to refresh the page.
Thanks for reporting!
Environment Details
Please indicate the following details about the environment in which you found the bug:
The positive constraints has the parameter called
strict_boundaries
, but in the examplestrict
is given, which results in the error: