Situation: have some monthly and weekly time series. Weekly data is fine. Took the monthly data and used "prepare time series data" application to spread the data down to the weekly level for each month. Output data set (weekly data) from that procedure looks okay, but is somewhat different than my first weekly data set in that the "date" variable has a date format "WEEKDATXw.d", whereas the first weekly data set date is formated "DATETIMEw.d" (a date/time format). (This is the default from the process.)
The informats for both appear to be the same - DATETIMEw.d. I decide "what, me worry? Enterprise Guide will know how to do the merge."
Result 1: Alas, it does not work. No error message in the log, but the resulting output data set is empty. Changing the type of join is not successful--I get either one or the other data set in the output, but not the combined series from both original sets.
So: I decide to go back to the dataset that resulted from the "prepare time series" process. I change the format to DATETIMEw.d on the off chance that that will create something that is "mergable". However, what I end up with after changing the format is a Date/time series that looks like every observation is from the same calendar day, with the only thing changing in each successive observation is the "second" increment. Obviously, this is NOT what is going on.
I am at a loss how the transformation process has caused this problem. Any suggestions before I go to Tech Support?
Thanks.