Displaying DataFrame with values from range [1e-7, 5e-7] shows 0 in some situations.

[tomazberisa]

Hi,

For illustration purpose, I'll assume the following dataset:

In [1]: d=pd.DataFrame({'col1':[9.999e-8, 1e-7, 1.0001e-7, 2e-7, 4.999e-7, 5e-7, 5.0001e-7, 6e-7]})

In [2]: d
Out[2]:
           col1
0  9.999000e-08
1  1.000000e-07
2  1.000100e-07
3  2.000000e-07
4  4.999000e-07
5  5.000000e-07
6  5.000100e-07
7  6.000000e-07

I've noticed the following behavior (in pandas 0.16 and 0.15.2):

1. When values from range [1e-7, 5e-7] are displayed along with values that are less than 1e-7, the output is OK:

In [3]: d[0:6]
Out[3]:
           col1
0  9.999000e-08
1  1.000000e-07
2  1.000100e-07
3  2.000000e-07
4  4.999000e-07
5  5.000000e-07

2. When values exclusively from that range are displayed, the output is 0:

In [4]: d[1:6]
Out[4]:
   col1
1     0
2     0
3     0
4     0
5     0

3. When values from that range are displayed along with values greater than 5e-7, the output for them is 0.000000:

In [5]: d[1:8]
Out[5]:
       col1
1  0.000000
2  0.000000
3  0.000000
4  0.000000
5  0.000000
6  0.000001
7  0.000001

I'm assuming it's a rounding / output formatting error.

Also:

In [6]: pd.show_versions()

INSTALLED VERSIONS
------------------
commit: None
python: 3.4.2.final.0
python-bits: 64
OS: Linux
OS-release: 2.6.32-431.el6.x86_64
machine: x86_64
processor: x86_64
byteorder: little
LC_ALL: None
LANG: en_US.UTF-8

pandas: 0.16.0
nose: 1.3.4
Cython: 0.21.2
numpy: 1.9.2
scipy: 0.14.0
statsmodels: 0.6.1
IPython: 2.3.1
sphinx: None
patsy: 0.3.0
dateutil: 2.4.1
pytz: 2015.2
bottleneck: None
tables: None
numexpr: None
matplotlib: 1.4.2
openpyxl: None
xlrd: None
xlwt: None
xlsxwriter: None
lxml: None
bs4: None
html5lib: None
httplib2: None
apiclient: None
sqlalchemy: None
pymysql: None
psycopg2: None

[jreback]

could be a bug or just an odd setting of options: http://pandas.pydata.org/pandas-docs/stable/options.html#list-of-options

want to see what is affecting the output here?

[tomazberisa]

I've looked at my options and all that are related to formatting floats are set to defaults (display.chop_threshold is None, display.float_format is None, display.precision is 7).

It seems to be an issue in pandas/core/format.py - working on a fix.