select and delete records older than 3 years

[JustinWolff]

We are attempting to maintain a dataset to include records only for the past 3 years.  New records are appended daily (via other scripted geoprocessing).

The table is in Enterprise SQL, and contains both varchar date (dd/mm/yyyy) and datetime (yyyy-mm-dd hh:mm:ss:ms) fields.

How do I select records that are 3 years older than yesterday (for deletion)?  I thought I was going down the correct path with this in SQL but I'm not getting far:

   DateTimeField < DATEADD(yy, -3, CONVERT(datetime, CONVERT(varchar, GETDATE(), 101)))

I need selections to the day, not the datetime (e.g. On March 11, the new records from March 10 are appended, and anything more than 3 years older than March 10 are deleted.  It's to the full date, not the actual time of day when the script is run.)  This selection will be used within a scheduled .py daily.

Thanks

Justin

[Kara_Shindle]

We use something like this where the days can be adjusted based on need for an extraction.

days = 20
"last_edited_date >= (CURRENT_Date-" + str(days) + ")

[JustinWolff]

Thanks Kara.  We're actually redesigning this project now.  I'm not sure if your suggestion would account for leap year days, etc.  Regardless, we're moving on for now but thanks again.