Squelch some setting with copy warnings #1100
Open
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
On "src/egon/data/datasets/pypsaeursec/__init__.py" and with version `23.1.0`. Using `--preview` first followed by a normal invocation pulls in new style fixes but also makes sure, that only those remain, which will not conflict with normal `black` usage, e.g. when pre-commit hooks are run. This one removes some extraneous blank lines and wraps overly long ones.
Take advantage of the fact that Python automatically joins adjacent strings by breaking the string at appropriate places to stay within the line length limit. Then wrap the two strings in parenthesis, making it one expression which can be spread over multiple lines. That way the strings can each be on individual lines and we're staying within 79 characters.
Actually, it's just to make the code shorter while also making it a bit more readable. (IMHO of course.)
Using `== False` can not only be unnecessarily verbose, but also surprisingly incorrect, as `0 == False` is `True`. Since `pandas` doesn't want us to use `not` with boolean `Series`, `~` is the correct choice for flipping those `Series`. This also squelches `flake8` complaints.
gnn
force-pushed
the
features/#1097-squelch-some-setting-with-copy-warnings
branch
from
aa8bed9
to
0b3c8dc
Compare
Replace single and dual letter variable names with descriptive ones. Remove comments as the variable names now serve a purpose comparable to that of the comments. Feel free to reintroduce the comments if you feel that they serve an additional informatory purpose. But please expand them a bit, at least noting that they are examples of what's happening in specific exemplary loop iterations. That is, if I understood them correctly. If I'm wrong, please expand them to prevent that misunderstanding.
The `component_t` object is an instance of `pypsa.descriptors.Dict`, which allows both, attribute access and key lookup. Therefore it's not necessary to use `getattr`.
The usage of `i` to store `DataFrame` column names was very confusing to me. It correctly suggested (to me) that `i` is an integer index so I got that `x["foo"][i]` is accessing column `"foo"` of the `DataFrame` stored in `x` and then pulling out the value at index `i` of the resulting `Series`. But since I now associated `i` with being an integer index, the left hand side of `x["foo"][i] = y[i].values.tolist()`, i.e. `y[i].values.tolist()`, didn't make sense to me, because `y[i]` looks like it's pulling out the single value of a `Series` or list at an integer index. But it doesn't. It actually pulls the column `i` out of the `DataFrame` `y`. That can then be converted to a list and is assigned to the `DataFrame` cell `x["foo"][i]`. Figuring this out took me hours, so I decided to change the variable name to `column` to convey that it's meant to store a column, even if it is sometimes used to index a row. What is indexed is then made explicit by using `.at` and `.loc` as appropriate. Using `.at[row, column]` also serves the purpose of squelching some instances of ``` SettingWithCopyWarning: A value is trying to be set on a copy of a slice from a DataFrame. Try using .loc[row_indexer,col_indexer] = value instead See the caveats in the documentation: https://pandas.pydata.org/pandas-doc# ``` Using `.at[r, c]` is necessary in order to tell `pandas` that we want to assign a sequence to a single `DataFrame` location. Using `.loc[r, c]` `pandas` would try to somehow align the sequence with the cell's neighbourhood which usually results in an error.
Instead of `[c]`. This hopefully squelches a few instances of ``` SettingWithCopyWarning: A value is trying to be set on a copy of a slice from a DataFrame. Try using .loc[row_indexer,col_indexer] = value instead See the caveats in the documentation: https://pandas.pydata.org/pandas-doc# ``` It should also make sure that the assignment always happens and isn't lost in an ephemeral copy.
At least in the places where I noticed. This makes it easier to see, whether the indexing is done into a dictionary, `Series` or list, i.e. a normally one dimensional structure, or a `DataFrame`, which is always two dimensional.
gnn
force-pushed
the
features/#1097-squelch-some-setting-with-copy-warnings
branch
from
0b3c8dc
to
e908207
Compare
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Fixes #1097.
Before merging into
dev-branch, please make sure thatCHANGELOG.rstwas updated.blackandisort.Dataset-version is updated when existing datasets are adjusted.continuous-integration/run-everything-over-the-weekend-2022-11-10-branch.Schleswig-Holsteinmode.Everythingmode.