inplace kwarg must be of bool type, but other boolean kwargs don't have this restriction

[colinmorris]

Code Sample, a copy-pastable example if possible

Using DataFrame.set_index as an example, but this applies to other methods that take an inplace kwarg:

# The following work
df.set_index('foo', drop=0)
df.set_index('foo', append='yes please')
df.set_index('foo', verify_integrity=22.7)
# This doesn't
df.set_index('foo', inplace=1)

The last line raises ValueError: For argument "inplace" expected type bool, received type int.1

Problem description

This behaviour seems inconsistent.

Expected Output

The last line should proceed successfully with inplace treated as True.

Or every boolean kwarg should be type-checked, though I like this option much less. I think using 0/1 as shorthand for True/False is a pretty common idiom.

It looks like this is the result of a somewhat recent 'bug fix': #14189. To me, it feels like a regression.

Output of pd.show_versions()

INSTALLED VERSIONS ------------------ commit: None python: 2.7.12.final.0 python-bits: 64 OS: Linux OS-release: 4.4.0-78-generic machine: x86_64 processor: x86_64 byteorder: little LC_ALL: None LANG: en_US.utf8 LOCALE: en_US.UTF-8

pandas: 0.20.1
pytest: None
pip: 9.0.1
setuptools: 36.0.1
Cython: None
numpy: 1.12.1
scipy: 0.19.0
xarray: None
IPython: 5.4.1
sphinx: None
patsy: None
dateutil: 2.6.0
pytz: 2017.2
blosc: None
bottleneck: None
tables: 3.3.0
numexpr: 2.6.2
feather: None
matplotlib: 2.0.2
openpyxl: None
xlrd: None
xlwt: None
xlsxwriter: None
lxml: None
bs4: 4.4.1
html5lib: 0.9999999
sqlalchemy: None
pymysql: None
psycopg2: None
jinja2: 2.9.6
s3fs: None
pandas_gbq: None
pandas_datareader: None

[gfyoung]

I think using 0/1 as shorthand for True/False is a pretty common idiom.

Perhaps when checking in if statements, but as parameters, I don't think so. I disagree that this is a regression, but I am +1 for checking all booleans in this function though.

[jreback]

We should fix all boolean input arguments with this check. We have in effect some strict type checking.

@colinmorris if you read the other issue, the problem is this can easily be from user error, where they are passing args and not as keyword arguments. So its nice if pandas can fail fast with an obvious error.

PR's to validate boolean input args welcomed!

[enriquebrgn]

I'll start working on this. Just a quick question, do I have to validate boolean input args just for the set_index function or also for every function with boolean kwargs?

[jreback]

ideally we should do this for all function args that are boolean

start with a certain class of them

e.g. like inplace (though these are mostly done i think)

[pambot]

@TomAugspurger [Scipy2017 Sprint] I decided the route to take is to type check using validate_bool_kwarg for all of the other kwargs, not just inplace by searching through the docstrings for all of the methods for boolean kwargs.

[gfyoung]

@pambot : That's going to be a little laborious, but if you can do that, (at least for some chunk of them), then go right ahead!