Power BI Dataflow with column type text having null value is rendered as blank in dataset

A bit late to the party, but..

When you ingest data from dataflows, you do indeed get blank values instead of 'null'-values. It's not a bug, as it is documented as an unsupported datatype. That's not clear when you are creating a dataflow, as then you can choose and see the nulls in your data. 

When you use a dataflow to write data to a datamart, and use that datamart to connect to Power BI, than you do get null-values. This has probably something to do with the way the data is stored: when you write to a datamart data is stored in an Azure SQL db, which supports null-values. When you only use the dataflow, data is stored in a text file on a blob-store. There a null value is just an empty string. 

Unfortunately that leads to only one solution when working directly with dataflows to Power BI: transform all the blank values to nulls when you need them. You can use a Power Query statement like this one to do that for an entire table at once: 

BlankToNull = Table.ReplaceValue(#"Previous Step", "" ,null,Replacer.ReplaceValue, Table.ColumnNames(#"Previous Step"))