DOC: update the pandas.DataFrame.clip_lower docstring #20289
There are no files selected for viewing
| Original file line number | Diff line number | Diff line change |
|---|---|---|
|
|
@@ -5716,24 +5716,86 @@ def clip_upper(self, threshold, axis=None, inplace=False): | |
|
|
||
| def clip_lower(self, threshold, axis=None, inplace=False): | ||
| """ | ||
| Return copy of the input with values below given value(s) trimmed. | ||
|
|
||
| Elements below the `threshold` will be changed to match the | ||
| `threshold` value(s). | ||
|
|
||
| Parameters | ||
| ---------- | ||
| threshold : float or array-like | ||
| Lower value(s) to which the input value(s) will be trimmed. | ||
| axis : {0 or 'index', 1 or 'columns', None}, default None | ||
| Align object with threshold along the given axis. | ||
| inplace : boolean, default False | ||
| Whether to perform the operation in place on the data. | ||
|
|
||
| .. versionadded:: 0.21.0 | ||
|
|
||
|
There was a problem hiding this comment. Sorry if I forgot some discussion about it, but don't we want a There was a problem hiding this comment. Yes good spot we definitely still want this - @adatasetaday can you add back in? I know we had some back and forth about |
||
| See Also | ||
| -------- | ||
| DataFrame.clip : General purpose method to trim `DataFrame` values to | ||
| given threshold(s) | ||
| DataFrame.clip_upper : Trim `DataFrame` values above given | ||
| threshold(s) | ||
| Series.clip : General purpose method to trim `Series` values to given | ||
| threshold(s) | ||
| Series.clip_lower : Trim `Series` values below given threshold(s) | ||
| Series.clip_upper : Trim `Series` values above given threshold(s) | ||
|
There was a problem hiding this comment. Ideally, for
So, the same in There was a problem hiding this comment. Only caveat here is that the user is working on a doc that is shared by I advised just removing it to not get hung up on it for now, the other option would be to explicitly list both There was a problem hiding this comment. Sorry, I was waiting for a little clarification, since I'm not sure how to proceed. I can go either way, but I haven't seen a conclusive definition. Was a decision made on how to proceed with these? There was a problem hiding this comment. If I'm not wrong, in another ticket the final implementation just listed everything. As @WillAyd says in some cases the same method being documented will be self-referenced, but it's not a big issue. |
||
|
|
||
| Returns | ||
| ------- | ||
| Series or DataFrame | ||
| Same type as caller. | ||
|
There was a problem hiding this comment. For this method may be it's a bit obvious, but for consistency I'd say in the description what this method returns (e.g. Original data with values trimmed). The note about the time is all right (not sure if other methods have it as part of the type or the description). There was a problem hiding this comment. Would any of these work? Series or DataFrame (same type as caller) Series or DataFrame There was a problem hiding this comment. They look great. If you want to take a look at the latest merged pull requests, I think this happens in few, and may be there is a way more standard than another. If you don't find any, from my side you can choose whatever you want. There was a problem hiding this comment. Based on what I see in the code, this seems to be the most common format type of caller I was thinking that there would be some value in creating PRs for standarizing these kind of things. For example, do one that fixes the axis parameters in ALL methods to match the format agreed in the discussions in this sprint. Another one could be done to make sure there is no period after the versionadded (which a lot of them have it, probably due to the validating script complaining if it was not there). Would this be a good idea? Who should I contact if I want to take care of some of these? There was a problem hiding this comment. Feel free to do it. I'd probably wait until most of the PRs from the sprint are closed, to avoid conflicts. And more than contacting someone, you can create an issue (or one per change) with what you're planning to do, you can get some discussion and feedback before doing all the work. |
||
|
|
||
| Examples | ||
|
There was a problem hiding this comment. Nice job here - I think the examples are very clear |
||
| -------- | ||
| >>> df = pd.DataFrame({'a': [0.740518, 0.450228, 0.710404, -0.771225], | ||
| ... 'b': [0.040507, -0.45121, 0.760925, 0.010624]}) | ||
|
There was a problem hiding this comment. Personally I find the data used in #20212 easier to understand. It requires a decent amount of concentration to see what the examples are doing with so many decimal numbers. |
||
|
|
||
| >>> df | ||
| a b | ||
| 0 0.740518 0.040507 | ||
| 1 0.450228 -0.451210 | ||
| 2 0.710404 0.760925 | ||
| 3 -0.771225 0.010624 | ||
|
|
||
| Clip to a scalar value | ||
|
|
||
| >>> df.clip_lower(0.2) | ||
| a b | ||
| 0 0.740518 0.200000 | ||
| 1 0.450228 0.200000 | ||
| 2 0.710404 0.760925 | ||
| 3 0.200000 0.200000 | ||
|
|
||
| Clip to an array along the index axis | ||
|
|
||
| >>> df.clip_lower([0.2, 0.4, 0.6, 0.8], axis=0) | ||
| a b | ||
| 0 0.740518 0.200000 | ||
| 1 0.450228 0.400000 | ||
| 2 0.710404 0.760925 | ||
| 3 0.800000 0.800000 | ||
|
|
||
| Clip to an array column the index axis | ||
|
There was a problem hiding this comment. I believe for consistency here you meant to say "Clip to an array along the column axis" |
||
|
|
||
| >>> df.clip_lower([0.5, 0.0], axis=1) | ||
| a b | ||
| 0 0.740518 0.040507 | ||
| 1 0.500000 0.000000 | ||
| 2 0.710404 0.760925 | ||
| 3 0.500000 0.010624 | ||
|
|
||
| Clip in place | ||
|
|
||
| >>> df.clip_lower(0.2, inplace=True) | ||
| >>> df | ||
| a b | ||
| 0 0.740518 0.200000 | ||
| 1 0.450228 0.200000 | ||
| 2 0.710404 0.760925 | ||
| 3 0.200000 0.200000 | ||
| """ | ||
| return self._clip_with_one_bound(threshold, method=self.ge, | ||
| axis=axis, inplace=inplace) | ||
|
|
||
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Let's use "caller" instead of "input". FWIW I guess it's not technically true that it returns a copy of the caller because
inplacewould modify directly. Any idea on how to make this statement more accurate?There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Something along the line of
Return a copy of the caller (or the caller itself if inplace == True) with values below given
thresholdtrimmed.?
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Hmm well I don't think you need to get in the discussion of return values since that's documented later in the docstring. First line is supposed to be very simple so I would mirror what's there for
clipand say something likeTrim the caller's values below a given thresholdThere was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Ok. Since Series.clip_lower uses the same docstring (the function is defined in generic.py) I thought that I should change the "See Also" description for Series.clip_lower to be the same. But when I tested the generation of the HTML I realized that if we use the "See Also" as it is right now, for Series.clip_lower we will not have a reference to DataFrame.clip_lower and it will have a reference to itself.
Does this make sense? Is there an elegant way to solve it?
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Yes that makes sense. Typically in this case you could use the
@Substitutiondecorator - you'll see some other functions using in that same module.That said, since your method is currently just inherited by
SeriesandDataFrameand not necessarily overriden, I don't think there's a good way without overriding those implementation (if even just to call super) to implement Substitution. Unless you can think of a better way, I'd suggest you remove theclip_lowerreference in See Also and open up a separate issue to try and make substitution workThere was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Nope, I can also only think of overriding in order for it to work. But is it worth the effort just to make the "See Also" complete? Would a reference to just clip_lower (without DataFrame. or Series. prefixes) be a good enough option?
I'll sleep through it to see if I can think of something better.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I would say just remove but if you come up with a better idea then by all means