One way is the window recipe:
In the post filter, you can add other conditions at will. In the output tab, you can remove the rownumber column if yo don't need it afterwards.
Depending on your use case, there are several possibilities. For a start, I would use a visual “group recipe”. It allows to keep only the line with the most recent timestamp.
- group by phone number
- in the recipe settings, on the line “phone number”, click “last, ordered by: <column containing time>”:
If you need to keep all lines, once you have selected the most recent phone number into a new dataset, you could also join this new with the original one: any line of the original dataset with a match in the new dataset is a correct phone number, any other line is to be flagged.
If you need to keep all lines, a window recipe should also do the trick, but this is more advanced.