Marcel Beugelsdijk on 10 Jan 2017 15:32:21
With the function Table.ExpandTableColumn, a column with nested tables can be expanded. However, the data types of the columns of the nested tables are not applied to the new columns after expansion.
This behaviour is different from Table.ExpandRecordColumn, where data types of the nested records are applied to the new columns after expansion.
So this should also be possible with Table.ExpandTableColumn.
This will also further improve the "combine binaries" functionality that was already strongly improved with the November 2016 update of Power BI Desktop:
currently, you still have to change data types after combining binaries, while these data types may have been explicitly defined in the sample query that is created from "combining binaries".
- Comments (2)
RE: Preserve data types when expanding table columns
This is really bad. It means every NestedJoin (which is default join from UX) messes up your column types on expansion. Every single one. Same issue when adding Index via GroupBy/ExpandTableColumn.And these errors are very hard to track down. Table.Combine can just hang spinning trying to reconcile two tables where in one a column is text and in another it's Any.So bad this is, I had to write a helper function to reapply column types based on previous types. Should be handled by the language itself.
RE: Preserve data types when expanding table columns
This may be related to: http://community.powerbi.com/t5/Issues/combine-binaries-on-csv-not-flowing-Type-Change-to-final-output/idc-p/111106#M1515