Data validation assignment

I am trying to use calculate_and_display_anomalies(statistics=X, schema=schema) while the statistics are generated using tfdv.generate_statistics_from_dataframe(Y, stats_options=options). The options are generated based on serving requirements.
When I try to compute anomalies, the results show no anomalies. Whereas, I should see that metformin-rosiglitazone, metformin-pioglitazone, payer_code and medical_specialty features have an anomaly (i.e. Unexpected string values) which is less than 1%.
Please let me know why I get the wrong output here. The only thing that is different from eval is the new statistics option I am using with serving df.

Hi @Asif_Mehmood
Try to compare the statistics generated using the evaluation data and serving data to see if there are any differences.
You can try to adjust the tfdv.Options passed to tfdv.generate_statistics_from_dataframe to better match the requirements of your serving data and see if that resolves the issue.

Muhammad John Abbas