Use arrow row format in GroupByHash rather than datafusion row format

**Is your feature request related to a problem or challenge? Please describe what you are trying to do.**
Arrow now has a row format https://docs.rs/arrow/26.0.0/arrow/row/index.html that is both well documented and likely to get additional functionality (as it will be used for sorting, etc)

It would be good to use that in DataFusion to benefit from upstream efforts including more type support as it is added

**Describe the solution you'd like**
It would be great to use the Arrow row format instead of `Compact` layout: https://github.com/apache/arrow-datafusion/blob/f2f846512ab032845de5dcee768a8a69ddf17eac/datafusion/row/src/layout.rs#L68

**Describe alternatives you've considered**
A clear and concise description of any alternative solutions or features you've considered.

**Additional context**
As part of https://github.com/apache/arrow-datafusion/issues/3940 and https://github.com/apache/arrow-datafusion/issues/3941 I believe @crepererum is working on this feature

cc @yjshen and @tustvold

Start by reading datafusion/row/src/layout.rs around line 68 and the GroupByHash implementation, then compare the existing Compact layout with Arrow's row module linked in the issue. Trace the relevant row-format tests or callers before changing anything. Done means GroupByHash uses Arrow's row format and the existing behavior remains covered by the test suite.

Written by the indexing model from the issue text.