Lenient Cube Concatenate #5392
Description
✨ Feature Request
This has been split out from #4446. While merge and concatenate are conceptually linked, the work can be entirely separated; indeed we could deliver one without the other. This allows for smaller parcels of work that are easier to find time for.
Extend the lenient behaviour that underpins the Lenient Cube Maths, introduced in Iris v3.0, to CubeList.concatenate().
From discussions around CubeList.merge(), we know there is community appetite for this improvement. It should be noted that while it will make the concatenation process easier, it will not be a panacea to all the problems caused by Cube differences.
Motivation
Iris is strict for obvious reasons regarding its treatment of metadata, but this is completely unhelpful when users require to combine multi-model data together in order to perform a common operation, such as a multi-model mean.
Due to the current metadata strictness of Iris, this makes it close to impossible or incredibly painful for users, as Iris forces the user to level the playing field with regards to their metadata before allowing them to proceed.
This is a burden on all our users, and that burden could be removed or significantly reduced by offering opt-in lenient concatenate() behaviours IMHO.
Remember to up-vote this issue if you want to see it happen! We definitely take votes into account, even when it takes a while to see a project through.