INPUT_DATA_VERY_LONG when append to Dataset

manuelberbig Registered Posts: 8


when I'm trying to append content to a Dataframe I get the warning message "INPUT_DATA_VERY_LONG occurred 22 times". The data contains strings as column values that can be very long. This warning results in a Dataset with columns that are completely empty. Some of them are fine. I used the option append to Dataset. Before the appending the columns were filled with values. After appending new content to the Dataset even the columns that were ok before are now NaN values. See the screenshot for more details. It is also not possible to scroll in this warning message. The Job doesn't fail, just the output dataset is broken.

Is this a Bug?

Kind Regards

Operating system used: Linux



  • Catalina
    Catalina Dataiker, Dataiku DSS Core Designer, Registered Posts: 135 Dataiker

    Hi @manuelberbig

    The warning message "INPUT_DATA_VERY_LONG" comes from having unusually large values for a row. This is usually caused by a misplaced quoting character or some other kind of malformed data.

    This warning might not be related to the broken output dataset.

    For further assistance I would suggest you open a support ticket on and provide the job diagnostic.

Setup Info
      Help me…