[new]: `is_table_row_number_anomalous(fully_qualified_table_name)` about bigfunctions HOT 2 OPEN

I can't understand clearly. Does that mean comparing today's row count with yesterday's and today's row count with the day before yesterday and so on up to 7 days latest days?

For example - if today is 28 April 2023

then the comparison will be

Is this need to be done or just have to compare it with 7 days old table? and output as true or false.

from bigfunctions.

I was thinking on:

• getting table row count for today using INFORMATION_SCHEMA
• getting table row count for latest 7 days using INFORMATION_SCHEMA and time travel
• calling prophet bigfunction to forecast the row count of today given the row counts of the latest 7 days
• return an anomaly score given the distance from the prediction and the real value. The score computation must be clever.

We can also avoid prophet and use a business rule.

from bigfunctions.