Code Sample, a copy-pastable example if possible

import pandas as pd

df = pd.DataFrame({'a': [1, 9, 8, 11, -1, 6, 4, 9, 0],
                   'b': [9, 1, 2, 5, 6, 0, 2, 3, 1],
                   'c': [2, 0, 19, 1, 3, 2, 3, 2, 3]})

dummy = df.nsmallest(n = 1, columns=df.columns)
print(dummy)

Problem description

In the code above, the result should give -1, 0, 0 and not -1, 4, 3 as we're asking for the smallest number column-wise. Right now it returns the row with the minimun only fot the first column.

Expected Output

 a  b  c
-1  0  0

Output of pd.show_versions()

INSTALLED VERSIONS ------------------ commit: None python: 3.6.2.final.0 python-bits: 64 OS: Linux OS-release: 4.13.5-1-ARCH machine: x86_64 processor: byteorder: little LC_ALL: None LANG: es_ES.utf8 LOCALE: es_ES.UTF-8 pandas: 0.20.3 pytest: 3.2.1 pip: 9.0.1 setuptools: 36.4.0 Cython: 0.26 numpy: 1.13.1 scipy: 0.19.1 xarray: None IPython: 6.1.0 sphinx: 1.5.5 patsy: 0.4.1 dateutil: 2.6.1 pytz: 2017.2 blosc: None bottleneck: 1.2.1 tables: 3.4.2 numexpr: 2.6.2 feather: None matplotlib: 2.0.2 openpyxl: 2.4.8 xlrd: 1.1.0 xlwt: 1.3.0 xlsxwriter: 0.9.8 lxml: 3.8.0 bs4: 4.6.0 html5lib: 0.9999999 sqlalchemy: 1.1.13 pymysql: 0.7.9.None psycopg2: None jinja2: 2.9.6 s3fs: None pandas_gbq: None pandas_datareader: 0.5.0 **Comment From: bobhaffner** I don't believe [DataFrame.nsmallest](https://pandas.pydata.org/pandas-docs/stable/generated/pandas.DataFrame.nsmallest.html) will provide the behavior you're looking for. The `column` arg is **Column name or names to order by** To illustrate 
 dummy = df.nsmallest(n = 1, columns=['c','a','b'])
 print(dummy)
   a  b  c
1  9  1  0
Maybe `df.min()` is what you're after? **Comment From: daviddoji** Thanks for your answer @bobhaffner. Obviously the nsmallest method is not doing what I need. The example above was just to illustrate the actual behaviour. Imagine now that you need to take the mean value between the 4 smallest numbers for each column of a Dataframe, e.g. df.nsmallest(n = 4, columns=df.columns).mean() The result given by nsmallest will not give the proper value unless, as you well say, other strategy is used. **Comment From: jreback** @daviddoji This is correct, nsmallest is the smallest in the column ordering breaking ties, you prob want ``df.min()`` as @bobhaffner indicates 
In [7]: df.nsmallest(1, columns=df.columns)
Out[7]: 
   a  b  c
4 -1  6  3

In [8]: df
Out[8]: 
    a  b   c
0   1  9   2
1   9  1   0
2   8  2  19
3  11  5   1
4  -1  6   3
5   6  0   2
6   4  2   3
7   9  3   2
8   0  1   3

In [9]: df.min()
Out[9]: 
a   -1
b    0
c    0
dtype: int64