This one we need to be careful about. I am unclear where the old values come from. I believe it is the official Pauling definition. It might be old but it is what people have been using all this while. This can potentially cause havoc since MP uses the electronegativity in generating the formula.

Yes can never be too careful when changing such core data recordings, looks like to me it might came from the same source (wikipedia) as used in the PR (not sure if wikipedia updated its recording at some point?):

Maybe a quick check would be to query for all MP formulas, rerun the Compositiion parsing and reduced formula nad compare the results? If there are no changes to formulas, we can potentially consider it. But again, this is something often used. E.g., matminer uses it as a feature in some instances.

Just to make sure we're on the same page, this PR tries to update the "electron affinity" not the "electronegativity" (X):

And indeed Composition uses X to sort the formula:

And AFAIK there's no places across pymatgen accessing this property:

And I didn't see any place in matminer trying to use it either

Ah my bad... I thought it was the electronegativity. Electron affinity is much less problematic since I don't think there are any critical use of it in normal operations.

No worries at all, never hurt to double check in any case :)