Large data model guidance: concatenated fields and compression

[sasdfasdfsad]

The PBI Tabular model uses column compression which benefits from having values that repeat as it can compress them down.

However, there does not seem to be a way to rebuild those values easily when you are on the page designer and trying to show them in columns, labels, filters.

For example IBAN codes.  Stored together they are unique and can't really compress, but if we store them in their individual parts, each column can compress down very nicely.

What's the best approach to deal with these scenarios?  If I need to display the entire number in a filter or column, but want to avoid storing it as a full string due to size?

[lbendlin]

This is similar to the discussion about storing datetime values in separate tables for the date and time part.  You will of course need a sacrificial lamb - an index column that actually ties them together. That column needs to be small (4 byte integer for example) because obviously it cannot be compressed at all.