Replacing a column with iloc replaces another column with same name

[mitar]

Code Sample, a copy-pastable example if possible

import pandas

a = pandas.DataFrame({'a': ['0'], 'b': ['str']})

print('---')
print(a)

a.iloc[:, 0] = [int(v) for v in a.iloc[:, 0]]

print('---')
print(a)

b = pandas.concat([a, pandas.DataFrame({'b': ['str2']})], axis=1)

print('---')
print(b)

b.iloc[:, 2] = ['str3']

print('---')
print(b)

Problem description

The issue seems to be that if there is a DataFrame with duplicate column names and mixed dtypes, if I try to replace one column with another value, using iloc, another column with same name is replaced as well.

Expected Output

The final print should be:

   a    b     b
0  0  str  str3

And not:

   a     b     b
0  0  str3  str3

It is interesting that if I change concat line to (see renaming of column b to c):

b = pandas.concat([a, pandas.DataFrame({'c': ['str2']})], axis=1)

Then the output is correctly:

   a    b     c
0  0  str  str3

Also, if a.iloc[:, 0] = [int(v) for v in a.iloc[:, 0]] is commented out, it works also correctly.

Moreover, the following also work correctly (see the change of 0 column index into [0] column index, and similar for 2 (this latter change is not really necessary to make it work)):

import pandas

a = pandas.DataFrame({'a': ['0'], 'b': ['str']})

print('---')
print(a)

a.iloc[:, [0]] = [int(v) for v in a.iloc[:, 0]]

print('---')
print(a)

b = pandas.concat([a, pandas.DataFrame({'b': ['str2']})], axis=1)

print('---')
print(b)

b.iloc[:, [2]] = ['str3']

print('---')
print(b)

Output of pd.show_versions()

INSTALLED VERSIONS

commit: None
python: 3.6.3.final.0
python-bits: 64
OS: Linux
OS-release: 4.13.0-46-generic
machine: x86_64
processor: x86_64
byteorder: little
LC_ALL: None
LANG: en_US.UTF-8
LOCALE: en_US.UTF-8

pandas: 0.23.3
pytest: None
pip: 18.0
setuptools: 40.0.0
Cython: None
numpy: 1.15.0
scipy: None
pyarrow: None
xarray: None
IPython: None
sphinx: None
patsy: None
dateutil: 2.7.3
pytz: 2018.5
blosc: None
bottleneck: None
tables: None
numexpr: None
feather: None
matplotlib: None
openpyxl: None
xlrd: None
xlwt: None
xlsxwriter: None
lxml: None
bs4: None
html5lib: None
sqlalchemy: None
pymysql: None
psycopg2: None
jinja2: None
s3fs: None
fastparquet: None
pandas_gbq: None
pandas_datareader: None

[WillAyd]

Strange indeed - investigation and PRs welcome!

[fjdiod]

Seems problem is with

pandas/pandas/core/indexing.py

Lines 526 to 548 in a6c7387

	def setter(item, v):
	s = self.obj[item]
	pi = plane_indexer[0] if lplane_indexer == 1 else plane_indexer
	# perform the equivalent of a setitem on the info axis
	# as we have a null slice or a slice with full bounds
	# which means essentially reassign to the columns of a
	# multi-dim object
	# GH6149 (null slice), GH10408 (full bounds)
	if (isinstance(pi, tuple) and
	all(com.is_null_slice(idx) or
	com.is_full_slice(idx, len(self.obj))
	for idx in pi)):
	s = v
	else:
	# set the item, possibly having a dtype change
	s._consolidate_inplace()
	s = s.copy()
	s._data = s._data.setitem(indexer=pi, value=v)
	s._maybe_update_cacher(clear=True)
	# reset the sliced object if unique
	self.obj[item] = s

as I understand item is a column name, so on line 527 we're selecting both columns as they have the same name