df.loc[:, 'col'] returning a view, but df.loc[df.index, 'col'] returning a copy

[STguerin]

Code Sample, a copy-pastable example if possible

x = pd.DataFrame(zip(range(4), range(4)), columns=['a', 'b'])
print(x)
   a  b
0  0  0
1  1  1
2  2  2
3  3  3

q = x.loc[:, 'a']
q += 2
print(x)
   a  b
0  2  0
1  3  1
2  4  2
3  5  3

x = pd.DataFrame(zip(range(4), range(4)), columns=['a', 'b'])
print(x)
   a  b
0  0  0
1  1  1
2  2  2
3  3  3

q = x.loc[x.index, 'a']
q += 2
print(x)
   a  b
0  0  0
1  1  1
2  2  2
3  3  3

Problem description

[df.loc[:, 'col'] returning a view, but df.loc[df.index, 'col'] returning a copy, intended? how can I make sure it is returning a copy?]

Expected Output

I thought .loc[] was always returning a copy

pd.show_versions()

INSTALLED VERSIONS ------------------ commit: None python: 3.5.2.final.0 python-bits: 64 OS: Windows OS-release: 7 machine: AMD64 processor: Intel64 Family 6 Model 94 Stepping 3, GenuineIntel byteorder: little LC_ALL: None LANG: None pandas: 0.18.1 nose: 1.3.7 pip: 8.1.2 setuptools: 27.2.0 Cython: 0.24.1 numpy: 1.11.1 scipy: 0.18.1 statsmodels: 0.6.1 xarray: None IPython: 5.1.0 sphinx: 1.4.6 patsy: 0.4.1 dateutil: 2.5.3 pytz: 2016.6.1 blosc: None bottleneck: 1.1.0 tables: 3.2.2 numexpr: 2.6.1 matplotlib: 1.5.3 openpyxl: 2.3.2 xlrd: 1.0.0 xlwt: 1.1.2 xlsxwriter: 0.9.3 lxml: 3.6.4 bs4: 4.5.1 html5lib: None httplib2: None apiclient: None sqlalchemy: 1.0.13 pymysql: None psycopg2: None jinja2: 2.8 boto: 2.42.0 pandas_datareader: None

[jreback]

this is as expected (see #6149), but see below.

df.loc[:, columns] is treated as df[columns] which may return a view.

df.loc[indexer, columns] also may return return a view, but almost always does not in practice.

yes if indexer is df.index we could treat this as the former situation (IOW the indexer is exactly equals to the index of the frame). (its just a indexer.equals(df.index) type of comparison.

So i'll mark this as a compat issue if you'd do a pull-request would be great. This is actually a very small change, see https://github.com/pandas-dev/pandas/blob/master/pandas/core/indexing.py#L503

[STguerin]

Thanks Jeff appreciate your insight on this.

[mroeschke]

This looks fixed in master. Could use a test

In [29]: pd.__version__
Out[29]: '1.1.0.dev0+1216.gd4d58f960'

In [30]: x = pd.DataFrame(zip(range(4), range(4)), columns=['a', 'b'])
    ...: print(x)
   a  b
0  0  0
1  1  1
2  2  2
3  3  3

In [31]: q = x.loc[:, 'a']
    ...: q += 2
    ...: print(x)
   a  b
0  0  0
1  1  1
2  2  2
3  3  3

[simonjayhawkins]

This looks fixed in master. Could use a test

The fix may have been #30501, which may need to be reverted. The tests added in #34996 may need to be removed and this issue reopened.

[jorisvandenbossche]

As noted in the PRs for 1.1.4 that reverted the fix for this, we should actually investigate what the expected behaviour is, because it's not clear that the test that was added in #34996 is actually correct (cc @jbrockmendel)