Open ktavabi opened 2 years ago
@ktavabi can you please validate if the behavior remains with the latest version 3.4.0
@ktavabi can you please validate if the behavior remains with the latest version
3.4.0
I was able to reproduce in 3.5.0
.
@ktavabi @fabclmnt the issue is in the documentation and in the code by @ktavabi.
The issue in the documentation:
profile = sample.profile_report(description=description, minimal=True)
The problem is that the field description
is nested under the category dataset
.
The issue with the code snippet:
prfl = pp.ProfileReport(edf.sample(frac=.05), **{"descriptions": description})
The field descriptions
(with an s
) is used to describe the columns of the dataset.
The proper behavior can be achieved using:
prfl = pp.ProfileReport(edf.sample(frac=.05), **{"dataset":{"description": description}})
alternatively:
prfl = pp.ProfileReport(edf.sample(frac=.05), dataset={"description": description})
The documentation page on the dataset metadata is up to date (https://pandas-profiling.ydata.ai/docs/master/pages/use_cases/metadata.html#dataset-metadata) so it is only about this section: https://pandas-profiling.ydata.ai/docs/master/pages/use_cases/big_data.html#sample-the-dataset
I will only modify the documentation.
Current Behaviour
When adding a description to the profiler configuration using
**{kwargs}
I am getting aExpected Behaviour
I am expecting a profile report with said description, as seen here
Data Description
I am reproducing the error with the example meteorite dataset
Code that reproduces the bug
pandas-profiling version
3.2.0
Dependencies
OS
macos
Checklist