I am running the sage notebook as an appliance in virtual box. I can import Numpy and matplotlib, but I can't import pandas. The error I get is that there is no such module. Is anybody else having this problem? How do I fix it?
It is possible to install pandas; however, it is not so simple as using easy_install.
The latest version of pandas (0.11) requires numpy 1.6.1 and sage is still using 1.5.1
There are also some optional dependencies (numepxr, PyTables, HDF5) that you probably want to have.
I don't know why sage lags behind in core packages such as numpy
As you can see in [trac ticket 11334](http://trac.sagemath.org/sage_trac/ticket/11334), numpy will be upgraded to 1.7.0 in Sage 5.10 (and as explained, it was not possible to upgrade to 1.6.x inbetween). If you need this version of numpy now, you can wait a few days that sage-5.9 is out an install 5.10beta0 from sources. Also, feel free to contribute ;)
You would have to install pandas first, I imagine; it's not included in the usual Sage distribution. Some discussion on this is in this sage-support thread. You should be able to install it from the command line.
sage: !easy_install pandas
or something like that.
But first you have to download the pandas source code and the easy install. How do I do this for Sage 5.5 running in Virtual Box?
I think @kcrisman meant sage: !easy_install pandas The program `easy_install` is already included in sage's python. And `pandas` will be downloaded and installed by `easy_install`. EDIT: Just tried installing it and it requires Numpy >= 1.6.1. Not really sure if it is safe to upgrade Numpy.
Hmm, good point. See http://trac.sagemath.org/sage_trac/ticket/11334 for why we haven't done this yet.
Asked: 2013-02-05 14:43:50 +0100
Seen: 1,991 times
Last updated: Apr 27 '13
Copyright Sage, 2010. Some rights reserved under creative commons license. Content on this site is licensed under a Creative Commons Attribution Share Alike 3.0 license.
