Dataiku keeps giving me errors when trying to export to Tableau Hyper File
trying to export a file. It keeps giving me errors like the one below. These are double and decimal numeric values that have not changed through my flow… if I erase column 21 it will say the same thing with column 20 etc etc.
I have gone back through the flow and tried to export the file and it works up to a point and then it starts giving me errors. I have not changed the data types at all.
It errors after I did a prepare step where I have 17 calculations that i have built. But the error is not on any of the columns that are calculations. The error it points to is data type of fields that were fine in the step before I did my calculations.
External code failed
Failed to export rows : <class 'ValueError'> : cannot safely convert passed user dtype of float64 for object dtyped data in column 21
Best Answer
-
I solved it myself. Dataiku for some reason does not like divisions by 0s. It will for some reason make changes to the columns of data that are used as the numerator or denominator. In order to solve for this you have to write the formula this way. val('metric1')/(val(metric2)+0.0000001).