Why blank column headers in a source file will break the time dimension mapping?

Hi community,

 

just something that intrigues and bothers me equally:

 

why when loading a file in Anaplan that has some blank column headers alongside the time dimension across columns (like this one)

AlejandroGomez_0-1632499408692.png

 

Anaplan will not recognise the time dimension as such, and the mapping tab for the time dimension will look like the one for a regular list? (like this:)

AlejandroGomez_1-1632499488688.png

 

Is there a purpose for this that I don't see? Else, could we avoid this from happening?  Would be nice to be able to ignore certain columns during the data mapping process as a workaround for this issue. There is already an idea exchange: https://community.anaplan.com/t5/Idea-Exchange/Ignoring-some-columns-when-mapping-timescale-through-column/idi-p/68334

 

Thanks,

Alex 

Answers

  • If time is mapped to columns and the headers for one or more unmapped column are empty then the time dimension mapping tab expects a manual mapping of numbered columns to values. It looks like the first two columns are neither mapped to anything nor marked as ignored in the main mapping tab. Try doing this and if necessary saving the import definition and re-entering the dialog to see if that changes things.

  • Thanks for your answer @ben_speight 

     

    You are right, some of the columns without column header are not mapped to any dimension of my target module, and the Time Dimension is mapped as "Column Headers". In this scenario, Anaplan won´t recognise the Time dimension and therefore instead of having the "specific time mapping tab" you would get the "generic" list mapping tab for time, which is a pain as Anaplan wont match the time periods automatically.

     

    So, for the sake of the example. Lets say that I have 2 dimension in my target module: Line Items and Time. I ma the second column to Line items and Time is mapped to column headers. Indeed, the first column of my source file will remain unmapped and will spoil the time dimension mapping tab.

     

    For the example above: Could you please share some more details on how to "mark [columns] as ignored in the main mapping tab"?

     

    Being able to "ignore" columns will solve this glitch for sure but I thought that was not possible in Anaplan, therefore the link i shared to the idea exchange requesting it.

     

    Also, to the first part of your comment: "If time is mapped to columns and the headers for one or more unmapped column are empty then the time dimension mapping tab expects a manual mapping of numbered columns to values." Yes, that is what I noted, but my question is: why is that? is that behaviour designed on purpose? or is it somehow unexpected? (to me it is more of a glitch as is a bit annoying and adds no value?)

     

    Thanks again,

    Alex 

  • You're right, you can't tell the import to ignore columns without manually mapping them - which prevents any kind of matching or custom date format parsing. I'd forgotten this but mis-remembered that you can sometimes ignore target dimensions in mixed-dimensionality situations.

    The 'manually map time column headers' function typically comes into play when you don't have a header row at all, and the columns are numbered rather than named. They are also numbered if the header cell is blank, or the header cell value is duplicated in another column so not unique. You only need one such unmapped column to trigger it - ideally it would give us more control over this, and the design assumption that all cells in the file are relevant to the import could also be challenged (either by marking specific columns as ignored, ignoring blank ones or specifying a range of column headers for the dimension). I think flipping to manual map to guard against having to deal with a set of unmatchable columns was a conscious design decision though, if a somewhat cautious one.