Forum Discussion
Dataflow and Dataset treating null differently
Here are the screenshots -
Original DB Connected dataset - the literal "(null)" in the power query was replaced as "(Blank)" in the tabular model -
And here is the dataset that connects to the dataflow - the literal "(blank)" in the power query has been replaced by an empty string.
Greg_Deckler Thanks for taking notice of this issue. Would you be able to help us out. We have separately raised this with support and are waiting to see how that proceeds.
Hi, did you receive any reply from support on this behaviour?
Yes, but not a favorable one. The ticket was escalated to the product team. PG team said that dataflow tables are stored in csv files and as such have no difference conceptually between empty and null. Null and empty values from source are both saved as empty values. They have suggested that I need to add a step to modify null values from the source to "null" strings.
That means I would need to do this for every table that has a text column where there is a potential that the column could have blank values for some of the records.
This is apparently still the case. Had a client that processed the same spreadsheet in dataflows vs desktop, with the same Mcode. Service processed as blanks, desktop returned null. To mirror the table in both, he had to processs for both null and blanks.
