I know there was an issue in June with the NULL field and not being calculated properly, but it seems to still exist, at least in the service.
I have the following M code in a table:
let Source = ARALTADR, #"Filtered Rows" = Table.SelectRows(Source, each ([Customer Number] = "BED100" or [Customer Number] = "BED200")), #"Removed Other Columns" = Table.SelectColumns(#"Filtered Rows",{"Customer Number", "Alternate Address", "Pool Facility Codes"}), #"Filtered out nulls" = Table.SelectRows(#"Removed Other Columns", each ([Pool Facility Codes] <> null)), #"Merged Queries" = Table.NestedJoin(#"Filtered out nulls", {"Pool Facility Codes"}, tblBBBPoolFacilityCodes, {"FacilityCode"}, "tblBBBPoolFacilityCodes", JoinKind.LeftAnti), #"Expanded tblBBBPoolFacilityCodes" = Table.ExpandTableColumn(#"Merged Queries", "tblBBBPoolFacilityCodes", {"FacilityCode"}, {"FacilityCode"}), #"Removed Other Columns1" = Table.SelectColumns(#"Expanded tblBBBPoolFacilityCodes",{"Customer Number", "Alternate Address", "Pool Facility Codes"}) in #"Removed Other Columns1"
The third row is removing nulls from a text field. In the desktop app (June 2019) it works fine, but when I publish to the service, that filter is being ignored and is returning all records.
They clearly show as null in the Power Query editor on the desktop.
I changed my query to this, which moves the null filtering after the join (less efficient, but wanted to test it)
let Source = ARALTADR, #"Filtered Rows" = Table.SelectRows(Source, each ([Customer Number] = "BED100" or [Customer Number] = "BED200")), #"Removed Other Columns" = Table.SelectColumns(#"Filtered Rows",{"Customer Number", "Alternate Address", "Pool Facility Codes"}), #"Merged Queries" = Table.NestedJoin(#"Removed Other Columns", {"Pool Facility Codes"}, tblBBBPoolFacilityCodes, {"FacilityCode"}, "tblBBBPoolFacilityCodes", JoinKind.LeftOuter), #"Expanded tblBBBPoolFacilityCodes" = Table.ExpandTableColumn(#"Merged Queries", "tblBBBPoolFacilityCodes", {"FacilityCode"}, {"FacilityCode"}), #"Filtered Rows1" = Table.SelectRows(#"Expanded tblBBBPoolFacilityCodes", each ([Pool Facility Codes] <> null) and ([FacilityCode] = null)), #"Removed Other Columns1" = Table.SelectColumns(#"Filtered Rows1",{"Customer Number", "Alternate Address", "Pool Facility Codes"}) in #"Removed Other Columns1"
And the service still ignores the <> null filter.
I finally applied the text type again to the Pool Facility Code field above and that caused it to work correctly in the service. I thought this was all resolved last month?