Hi Christian
Edit distance functions are more suitable for text strings such as names and not really usable for dates.
For instance, 30-Nov-2024 and 1-Dec-2024 is only 1 day different but the strings are vastly different.
Regards, Peter
Hi Peter, thank you.
I found this solution as well and simply casted the date into varchars and it worked better.
I still have to adjust my rule but it's now a bit better.
If I had to compare date, would it be better to do something like a datediff with a range of days?
Thanks,
Christian
You could use a date range function but ensure it is combined with the matching of other attributes that can identify duplicates or similar records.
Christian Khun
Hi,
We're trying to match event data according to dates that might be roughly 1-2 days different according to our providers.
is it possible to create a matching rule using the SEM_EDIT_DISTANCE_SIMILARITY function?
Thanks,
Christian