1 Reply Latest reply on Oct 18, 2018 9:26 AM by patrick.byrne.0

    LOD calculation results in "View Data" don't match pane value

    Ryan Blosser

      Hello,

       

      I inherited a workbook that uses a bit of pre-Tableau data manipulation to normalize some dollar values in duplicate rows.  Basically, a "record" would be in 4 rows with $2000 dollars, and the user made each row say .25 (1/4) and $500 (1/2000).  Then if you sum them up, that's 1 record for $2000.  I wanted to replace this datasource so as not to really disturb any of the other formulas, etc on the workbook, and I figured that I could use something like 1/ {FIXED [id]: sum([dummy]) } , where dummy column = 1 for every row in the data.  Then I multiply that by the full dollar value row and sum those up.   So the results in View Data (the csv) are:

      A 1

      B 0.5

      B 0.5

      C 0.33

      C 0.33

      C 0.33

       

      and that summed up equals 3 which is good, that's what I want, I have 3 "real records".  That is not the result in a pane that has these all in the same level of detail though.  The answer I get is 1.83, which would be one row from each.  Because there are many views on the dash that have different levels of granularity, I didn't know if I could simply do a max(dollar value), because if a "record" (like B or C) spans across multiple members of a dimension (there is a reason why these rows are duplicated, after all), I would get a HIGHER number than the intended result, instead of a LOWER number than I am now.  Am I stuck here without a quick Tableau solution?  I really want a 1-dimension COUNTIF that is filter agnostic, there's no real gotcha here, it was just a matter of saving time with the current Excel manipulation.

       

      Thank you