I created a dataset from a csv in a datalake storage.
When I preview the data, via the dataset, every column shows correctly.
When I use the same dataset in a source transformation of a dataflow it shows the date columns as null
This happens despite that the projection tab of the dataflow manages to detect correctly the data types when I press the detect data type button
Any ideas why this might happen?
For some reason I had to specify the format in the Projection tab.
Then, when I previewed the data, it looked as expected: