2 Replies Latest reply on Jul 6, 2016 6:57 AM by Leslie Foster

    Mimic SQL Group By, for Standard Deviation

    Leslie Foster


      I have data that has a Date Time field, plus additional fields.  I wish to group the data such that the hour-of-day ('hour') part of the date field, and another, 'User' field, can be used to make total record counts.  From those counts, I wish to find averages (average number of records/user/hour-of-day).  Then I wish to also find the variability across users (standard deviation).


      I can pretty easily make histograms dividing up the record counts by hour or user, or both.  But I want to consider the counts in these subdivisions as values to be compared to each other.  I think this would mimic a "GROUP BY" and "SUM" in SQL.  The simplistic approach (by way of example) is to naively drag average for one of these fields, but of course that gives me a '1' for everything, since each record contributes 1 to the "Number of Records".  Also, it would be possible to make another data source, that "rolls up" these records, but hopefully, Tableau can do something better than that.


      Has anyone done something like this?


      Thanks, in Advance