BUG: df.clip does not handle inverted lower/upper bounds consistently

[jkew]

Pandas version checks

• I have checked that this issue has not already been reported.

• I have confirmed this bug exists on the latest version of pandas.

• I have confirmed this bug exists on the main branch of pandas.

Reproducible Example

import pandas as pd
import numpy as np
data = {'A': [-2, -1, 0, 1, 2], 'B': [-2, 1, 0, 1, 2]}
df = pd.DataFrame(data)
df
	A	B
0	-2	-2
1	-1	1
2	0	0
3	1	1
4	2	2
df.clip(lower=10, upper=[1,1])
A	B
0	10	10
1	10	10
2	10	10
3	10	10
4	1	1

Issue Description

This is undocumented behavior; but when upper is < lower, and upper is not a scalar the resulting data frame is not consistently producing the same results. If lower is an array, and upper is a scalar the values are fixed at whatever the scalar upper is.

I'm not sure if pandas should simply throw an error when the bounds are inverted, or what - but some consistency would be helpful.

Expected Behavior

Either throw an error when upper < lower or have a consistent way to handle the result between scalar and list-like things.

Installed Versions

INSTALLED VERSIONS

commit : 2e218d1
python : 3.10.9.final.0
python-bits : 64
OS : Darwin
OS-release : 21.4.0
Version : Darwin Kernel Version 21.4.0: Mon Feb 21 20:35:58 PST 2022; root:xnu-8020.101.4~2/RELEASE_ARM64_T6000
machine : arm64
processor : arm
byteorder : little
LC_ALL : None
LANG : en_US.UTF-8
LOCALE : en_US.UTF-8

pandas : 1.5.3
numpy : 1.24.2
pytz : 2022.7.1
dateutil : 2.8.2
setuptools : 67.6.0
pip : 23.0
Cython : None
pytest : 7.2.2
hypothesis : None
sphinx : None
blosc : None
feather : None
xlsxwriter : None
lxml.etree : None
html5lib : None
pymysql : None
psycopg2 : None
jinja2 : 3.1.2
IPython : 8.10.0
pandas_datareader: None
bs4 : 4.11.2
bottleneck : None
brotli :
fastparquet : None
fsspec : 2023.3.0
gcsfs : None
matplotlib : 3.7.1
numba : None
numexpr : None
odfpy : None
openpyxl : None
pandas_gbq : None
pyarrow : 10.0.1
pyreadstat : None
pyxlsb : None
s3fs : None
scipy : 1.10.1
snappy : None
sqlalchemy : 2.0.4
tables : None
tabulate : None
xarray : None
xlrd : None
xlwt : None
zstandard : 0.19.0
tzdata : None

[m-ganko]

take

[m-ganko]

Ok, I tested clip method and it is quite a mess.

>>> import pandas as pd
>>> import numpy as np

>>> data = {'A': range(10), 'B': range(10)}
>>> df = pd.DataFrame(data)
>>> df
   A  B
0  0  0
1  1  1
2  2  2
3  3  3
4  4  4
5  5  5
6  6  6
7  7  7
8  8  8
9  9  9

When lower < upper it works as expected:

>>> expected = df.clip(lower=3, upper=7)
>>> expected
   A  B
0  3  3
1  3  3
2  3  3
3  3  3
4  4  4
5  5  5
6  6  6
7  7  7
8  7  7
9  7  7
>>> all(expected == df.clip(lower=3, upper=[7, 7]))
True
>>> all(expected == df.clip(lower=[3, 3], upper=7))
True

When we swap lower and upper, we get 3 different results.

>>> df.clip(lower=7, upper=[3, 3])  # 1
   A  B
0  7  7
1  7  7
2  7  7
3  7  7
4  3  3
5  3  3
6  3  3
7  3  3
8  3  3
9  3  3
>>> df.clip(lower=[7, 7], upper=3)  # 2
   A  B
0  3  3
1  3  3
2  3  3
3  3  3
4  3  3
5  3  3
6  3  3
7  3  3
8  3  3
9  3  3
>>> df.clip(lower=7, upper=3)  # 3
   A  B
0  3  3
1  3  3
2  3  3
3  3  3
4  4  4
5  5  5
6  6  6
7  7  7
8  7  7
9  7  7

• Case # 1 - for this one I don't have explanation, it is wrong result
• Case # 2 - this one works same as numpy.clip, when lower > upper all values are equal to upper.
• Case # 3 - here lower and upper are swapped, but this behavior is not documented.

In my opinion, there are two reasonable approaches:

1. Follow numpy approach (# 2).
2. Raise exception when lower > upper.

I will start coding first one, to be consistent with numpy. If you have other idea, discussion is welcome.