Grouping recipe forcing extra column
I am trying to set up a grouping recipe but running it kept providing me with an error stating the number of columns did not match. After some investigation I've found out that the recipe is trying to add another column, that is identical to the partitioning column of the output dataset. It is stating it has only 3 output columns but it only runs if the output dataset has 4 columns (and only runs after ignoring a schema inconsistency warning).
See the picture attached, when running the grouping recipe on the left it first gives a warning that the output dataset has 4 columns and the recipe only has 3. When running however, it is able to provide us with the output dataset on the right, with a seemingly unexpected duplicate of the partitioning dataset "R_month".
Does anybody know why the grouping recipe is behaving in this way? Or how to disable it?
Operating system used: Windows
Answers
-
JordanB Dataiker, Dataiku DSS Core Designer, Dataiku DSS Adv Designer, Registered Posts: 296 Dataiker
Hi @SanderVW
,Would you please try reloading the input dataset by going to the dataset settings > schema > select "Check Now" > select "Reload schema from detected data". Then, try running the recipe once more.
If that does not solve the issue, would you please provide a screenshot of the error message that you mentioned?
Thanks!
Jordan