Pandas changes default numpy array format

Question

Is there any way to prevent pandas from changing the default print format for numpy arrays?

With plain numpy, I get:

>>> numpy.array([123.45, 0.06])
array([  1.23450000e+02,   6.00000000e-02])

After I import pandas, I get:

>>> numpy.array([123.45, 0.06])
array([ 123.45,    0.06])

Can I stop it from doing this as a configuration setting? I don't want to have to wrap every "import pandas" with a "foo=np.get_printoptions(); import pandas; np.set_printoptions(**foo)", but that's the best I can come up with.

As it is, if I import pandas in one place, I get doctest errors from another.

sbraden · Accepted Answer · 2012-06-26 05:28:27Z

2

You should only have to wrap the first import pandas with np.set_printoptions(foo) since python will cache it. See below:

import numpy
>>> numpy.array([123.45, 0.06])
array([  1.23450000e+02,   6.00000000e-02])
>>> import pandas
>>> numpy.array([123.45, 0.06])
array([ 123.45,    0.06])
>>> numpy.set_printoptions(edgeitems=3,infstr='inf', linewidth=75, nanstr='nan', precision=8, suppress=False, threshold=1000)
>>> numpy.array([123.45, 0.06])
array([  1.23450000e+02,   6.00000000e-02])
>>> import pandas
>>> numpy.array([123.45, 0.06])
array([  1.23450000e+02,   6.00000000e-02])

answered Jun 26, 2012 at 5:28

sbraden

4994 silver badges5 bronze badges

Sign up to request clarification or add additional context in comments.

1 Comment

Johann Hibschman Over a year ago

Thanks, but we're using nose to load in the tests, which means I don't know in what order the modules will be loaded. I could always put it in the top package-level init, but that feels invasive. Plus, I'd have to change everyone else's doctests. I can't help but think that pandas shouldn't be changing numpy package-level globals.

Wes McKinney · Accepted Answer · 2012-09-08 23:02:53Z

1

This is a bit irritating as I think that NumPy should have suppress=True as its default. I'm changing pandas to not change it.

answered Sep 8, 2012 at 23:02

Wes McKinney

106k32 gold badges146 silver badges109 bronze badges

1 Comment

Johann Hibschman Over a year ago

Thanks, Wes! That would be great. If I run into you at DataGotham, I'll say thanks in person.

Collectives™ on Stack Overflow

Pandas changes default numpy array format

2 Answers 2

1 Comment

1 Comment

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

2 Answers 2

1 Comment

1 Comment

Your Answer

Sign up or log in

Post as a guest

Related