TYP: pd.eval, DataFrame.(eval, query) #38625
Conversation
arw2019 added Typing 
pandas/core/computation/eval.py Outdated
| def eval( | ||
| expr, | ||
| parser="pandas", | ||
| parser: str = "pandas", |
There was a problem hiding this comment.
@simonjayhawkins IIUC we can't use Literal yet. what if we used enums as a stopgap
class ParserStr(enum): pandas = "pandas" numexpr = "numexpr" that way we could start filling these in where appropriate and when we can use Literal swapping them out will be easy (wont require hunting them down)
There was a problem hiding this comment.
i think we are now using Literal
pandas/core/frame.py Outdated
| ------- | ||
| DataFrame or None | ||
| DataFrame resulting from the provided query expression or | ||
| ndarray, scalar, pandas object, or None |
There was a problem hiding this comment.
"pandas object" means "Series or DataFrame" here? could confuse with PandasObject
pandas/core/frame.py Outdated
| DataFrame or None | ||
| DataFrame resulting from the provided query expression or | ||
| ndarray, scalar, pandas object, or None | ||
| The result of evaluation of the provided query expression or |
There was a problem hiding this comment.
"of evaluation of" -> "of evaluating"?
| @arw2019 can you address @jbrockmendel comments. |
| Needs rebase and a couple of comments are outstanding, but otherwise I think this is pretty close. |
| @jbrockmendel over to you. |
There was a problem hiding this comment.
There's no default arguments preceding inplace in either of these functions, so they should be straightforward to overload - this would obviate the need for the casts you've added (no objection to this being done in a follow-up PR though)
Also, I think @jbrockmendel 's commented on having a preference for using # type: ignore rather than cast, especially if it's for something which can be fixed later (e.g. here it can be fixed with overloads)
| @MarcoGorelli feel free to push or otherwise close. I was going to close as stale but @jbrockmendel commented it was pretty close. #38625 (comment) |
| PythonFuncType = Callable[[Any], Any] | ||
| | ||
| IndexableEvalResult = Union[ArrayLike, FrameOrSeriesUnion] | ||
| EvalResult = Union[IndexableEvalResult, Scalar] |
There was a problem hiding this comment.
is this meaningfully distinct from Any?
| closing as stale. |
4 participants
black pandasgit diff upstream/master -u -- "*.py" | flake8 --diffxref #38416 (comment)