Prevent Y / N column from being Auto-Detected as Boolean column

pyousefi

For my dataset I have over 200 columns with many of them Y/N flag columns. I would like to keep them as text and not infer the type.

There is a checkbox that I am leaving unchecked: Normalize booleans Normalize all possible boolean values (0, 1, yes, no, …) to 'true' and 'false'

Clément_Stenac

It's not possible to disable meaning inference. Note that it only affects the display meaning, it does not actually change anything in the way the data is stored. See here: https://doc.dataiku.com/dss/latest/schemas/definitions.html for more information about meanings vs storage types.

You can force the meaning to "Text" (the 'default' meaning) on multiple columns (mass actions) on the schema screen in the dataset settings.

pyousefi

Thanks Clement. The Mass Action is what I did after filtering for boolean. When I download the dataset after inference, those columns that have been inferred, are replaced with true / false.

Bayardo

This is one of the most annoying things so far about Dataiku. I would not care if it changed it to boolean true false if it left the Yes and No. Instead, it converts it to True/False.