TYP: define Index.any, Index.all non-dynamically #36929
Conversation
jreback
added
the
Numeric Operations
|
can you rebase |
| False | ||
| """ | ||
| # FIXME: docstr inaccurate, args/kwargs not passed | ||
| self._maybe_disable_logical_methods("any") |
There was a problem hiding this comment.
do we have a numpy validation routine for this and for all? eg. https://github.com/pandas-dev/pandas/blob/master/pandas/compat/numpy/function.py
…p-31160-index-logical
| or is_float_dtype(self.dtype) | ||
| ): | ||
| # This call will raise | ||
| make_invalid_op(opname)(self) |
There was a problem hiding this comment.
could this be done on the subclasses. more repetitive but would be easier to follow I think.
although, does disabling on subclasses obey LSP?
maybe a mixin that classes can opt into, to ensure that the base class or other code does not rely on any/all being available.
This would mean a base class for Index, to allow any/all on object Indexes?
There was a problem hiding this comment.
ATM we disable on subclasses (though by calling cls.add_disabled_whatever()), and i thought it was clearer to do this in just one place, but if people disagree im fine with going the other direction.
I don't think it's that big a deal because I expect to change it before long since, as mentioned in the OP, i think several of the currently-disabled methods should be enabled.
There was a problem hiding this comment.
I'm OK with perhaps revisiting after some of the currently-disabled methods are enabled.
@jbrockmendel maybe open issues for these and @jreback comment #36929 (comment) |
|
Thanks @jbrockmendel |
black pandasgit diff upstream/master -u -- "*.py" | flake8 --diffxref #31160, cc @simonjayhawkins
This unearths 2.5 issues:
1b) the RangeIndex methods don't have args/kwargs at all
self.categoriessupports it