How do I filter by most recent date?

damen

We have a dataset that get multiple 'loan_id' entries recorded (shown below)

However, when I go to remove duplicates on any of the columns, I generally get the wrong loan amounts.

The correct loan amounts are always the ones with the latest 'time_stamp'

I am hoping to remove any duplicate 'loan_id' based off the latest time stamp.

How to I do this? Is this possible using an ETL?

MarkSnodgrass

You can do this in Magic ETL with the Rank & Window and Filter tiles. In the Rank & Window tile, select the Rank function and order by time_stamp descending and partition by load_id. Add a filter tile after this and filter to where rank equals 1. This will give you just one entry per loan_id and it will be the one with the latest timestamp.

MarkSnodgrass

You can do this in Magic ETL with the Rank & Window and Filter tiles. In the Rank & Window tile, select the Rank function and order by time_stamp descending and partition by load_id. Add a filter tile after this and filter to where rank equals 1. This will give you just one entry per loan_id and it will be the one with the latest timestamp.

damen

@MarkSnodgrass ah those trainings are coming back now. much appreciated, sir!