Support remaining `Expr.is_<...>` mask methods

[dangotbanned]

Description

We have almost all of these supported, and completing that group seems pretty realistic to me.

They might all be possible with a default implementation in CompliantColumn (#2967) that just __invert__s another method?

is_not_null is the only one I'd push for being directly implemented for each backend for performance (see discord comment)

Overview

Method	Supported	Related
is_between	✅	(#609), (#1672)
is_close	✅	(#2962)
is_duplicated	✅	(#187)
is_finite	✅	(#1341), (#1825)
is_first_distinct	✅	(#187)
is_in	✅
is_infinite	❌	(#2962 (comment))
is_last_distinct	✅	(#187)
is_nan	✅	(#1625)
is_not_nan	❌
is_not_null	❌	(discord comment)
is_null	✅
is_unique	✅	(#187)

[danielgafni]

Can we also have is_not?

[dangotbanned]

Can we also have is_not?

Unless I'm misunderstanding something, is_not is not 😉 a method for polars.Expr

Did you mean polars.Expr.not_?

[danielgafni]

Oh, sorry, yes.

[dangotbanned]

Ah that would be part of

• Support operator aliases in Expr #2213