ENH: add DataFrame.is_unique method #37565
Conversation
topper-123
force-pushed
the
dataframe_is_unique
branch
2 times, most recently
from
c8c6064 to
1683699
Compare
pandas/core/base.py
Outdated
| @@ -1038,16 +1038,19 @@ def nunique(self, dropna: bool = True) -> int: | |||
| n -= 1 | |||
| return n | |||
|
|
|||
| def _is_unique(self) -> bool: | |||
There was a problem hiding this comment.
i don't think this necessary
pandas/core/frame.py
Outdated
| return self.loc[:, subset].is_unique() | ||
|
|
||
| if len(self.columns): | ||
| return self.apply(Series._is_unique) |
There was a problem hiding this comment.
this should just work with
return self.apply(lambda x: x.is_unique())
There was a problem hiding this comment.
Seemed to be the the same, but I guess someone could have subclassed _constructor_sliced, which could make a difference.
pandas/core/frame.py
Outdated
| DataFrame.duplicated : Indicate duplicate rows. | ||
| """ | ||
| if subset is not None: | ||
| subset = subset if is_list_like(subset) else [subset] |
| self = self[subset] | ||
|
|
||
| if len(self.columns): | ||
| return self.apply(lambda x: x.is_unique) |
There was a problem hiding this comment.
why does this not work on empties?
There was a problem hiding this comment.
A dataframe with no columns doesn't call the inner func in apply (here lambda x: x.is_unique), so the output (empty) series has to guess its own dtype. It guesses float64, which is wrong, so we have to special-case it.
|
For a DataFrame, |
|
@jorisvandenbossche that's true, and could be fixes by having an |
topper-123
force-pushed
the
dataframe_is_unique
branch
from
d411050 to
9f51784
Compare
|
This pull request is stale because it has been open for thirty days with no activity. Please update or respond to this comment if you're still interested in working on this. |
|
@topper-123 this was close, can you merge master and move the note to 1.3 |
topper-123
force-pushed
the
dataframe_is_unique
branch
from
9f51784 to
45f5b88
Compare
topper-123
force-pushed
the
dataframe_is_unique
branch
from
45f5b88 to
63643a0
Compare
|
Ok, done. |
|
Looks pretty good overall. Should there be a corresponding |
|
Thanks for the PR, but it appears that it's gone stale. Please let us know if you'd like to continue by merging master and targeting for 1.4. |
Currently, to see columns with unique values, we have to do
df.apply(lambda x: x.is_unique), which is a bit awkward.IMO checking for unique columns in a DataFrame is a common enough need to justify a method for that specific purpose.
Xref: #11948