I’d like to create a filter for my Q topic which excludes all blank values for a given field.
The dataset/source referenced by the topic is a csv file in s3, and ‘blank’ values in this context means cells with no value, as opposed to ‘null’ values or similar.
When authoring analyses this is quite simple to do, you create a filter of type ‘filter list’, chose ‘exclude’ for the rule and then from the list, select the blank value, like so:
When following this same procedure in the Q console (topic > data > add filter), saving the filter fails with the error “Something went wrong while updating”. The network request returns a 400 with the error message “InvalidInputException”.
Being able to apply a blank value filter to a topic field (rather than the dataset) would hugely improve UX for our topics, since in our implementation we rely on a single large denormalized dataset with many blank rows. Is it intended that this is not possible, or is this a bug? If it’s a bug I’ll head over to AWS support.
I tried it with other values not null or blank and it worked seamlessly as expected and described by you. Does it work for you with other data value lets say ABCya! ? just wana know if the issue is with null only. May be its temporary issue could you try again ?
Hi Deep! Thanks for the quick response. You’re correct, values which are not blank work fine. I only encounter the error when trying to exclude blank values.
Still possible to reproduce this today. Let me know if you need any additional information.
Is it possible for you to eliminate those null or empty in the data set itself? could you share possible sample csv as well if it does not contain and sensitive data ?
Hi Deep, unfortunately not due to the constraints of our implementation.
I’ve worked around this by creating a calculated field which replaces blank values for a given field with the string “blank”, then creating a filter for that field which excludes this string.
Very doable as a workaround, but seems inconsistent that you can’t create a topic filter for blank values while you can for analyses.
In any case thanks for your help and I’ll mark this as the solution if anyone encounters the same issue.
I understand the constraints. However, I am glad that you have workaround and thats what i wanted to recommend as well. Nevertheless, will keep a close eye on the subject and share with you if any update comes from fellow colleagues.