Date definition in .csv other then in other data sources

[kdemulder]

Hey,

Ok, disregard my previous post "Catching orphans". I found out where the orphans came from and solved all of them... Except one dimension keeps having some. But I found out what triggered this. A .csv-file that is generated by a server every week and that I am using. Now when I add this source and name the date-column the same way as in my other sources, it errors when I create the cube (Date invalid). So I went into that column's properties in transformer and stated the date input format. Now the cube builds without errors BUT... the dates come as orphans in the date dimensions and in a completely other format (these are shown in dd/mm/yyyy format whereas all other dates ar stated in yyymmdd format).
Does anyone have any suggestions as to what more I can try here?
Thx in advance!

 

[drlex]

You should be able to set the order of the date input in the drop down list for 'Date input format' under the 'time' tab of the properties window of the relevant .csv field. Works for me.
lex

soi la, soi carre

 

[kdemulder]

Yeah, indeed, that's what I did and that took care of the error, but now these dates come as orphans as where the other dates (from other sources) come through in the correct categories (as in weeks, months etc).

(The data in this .csv does not span the entire period, only a couple of weeks of the total amount of dates that are present in the cube.)

 

[drlex]

Sorry - misread your post. Would the 'show scope' facility on the masures and data sources help identify the problem?

soi la, soi carre

 

[DoubleD]

If you've named the date field the same in all datasources, and defined it as a date field, and set the input date format correctly for the CSV they should all be formatted identically in the cube.

Pain is stress leaving the body.

DoubleD