I run a SAS program that computes a new data series from high frequency financial data and I want to aggregate it to different intervals, using the 'interval=..' option in the 'ID'-statement of the PROC Timeseries procedure.
When doing this for MIN or HOUR, it all works fine, but when I use DAY as aggregation level, the following error message is returned:
"ERROR: The ID variable value ObsDT=03JAN07:08 is extreme and invalid at observation number 2 in data set WORK.TEMP"
The strange thing is that it does work when aggregating to HOUR24, which should be the same as DAY as far as I know.
The variable ObsDT has the 'datetime.' format and the value 03JAN07:08:30:13, which is the same value as the previous and next observation. Could the fact that multiple observations have the same value have something to do with the error message?