Installation¶
You have the option to install an official release or to build the development version. If you choose to install from source and are running Windows, you will have to ensure that you have a compatible C compiler (MinGW or Visual Studio) installed. How-to install MinGW on Windows
Python version support¶
Officially Python 2.6 to 2.7 and Python 3.1+, although Python 3 support is less well tested. Python 2.4 support is being phased out since the userbase has shrunk significantly. Continuing Python 2.4 support will require either monetary development support or someone contributing to the project to maintain compatibility.
Binary installers¶
All platforms¶
Stable installers available on PyPI
Preliminary builds and installers on the Pandas download page .
Overview¶
Platform | Distribution | Status | Download / Repository Link | Install method |
---|---|---|---|---|
Windows | all | stable | All platforms | pip install pandas |
Mac | all | stable | All platforms | pip install pandas |
Linux | Debian | stable | official Debian repository | sudo apt-get install python-pandas |
Linux | Debian & Ubuntu | unstable (latest packages) | NeuroDebian | sudo apt-get install python-pandas |
Linux | Ubuntu | stable | official Ubuntu repository | sudo apt-get install python-pandas |
Linux | Ubuntu | unstable (daily builds) | PythonXY PPA; activate by: sudo add-apt-repository ppa:pythonxy/pythonxy-devel && sudo apt-get update | sudo apt-get install python-pandas |
Linux | OpenSuse & Fedora | stable | OpenSuse Repository | zypper in python-pandas |
Dependencies¶
NumPy: 1.6.1 or higher
python-dateutil 1.5
- pytz
- Needed for time zone support
Recommended Dependencies¶
- numexpr: for accelerating certain numerical operations. numexpr uses multiple cores as well as smart chunking and caching to achieve large speedups.
- bottleneck: for accelerating certain types of nan evaluations. bottleneck uses specialized cython routines to achieve large speedups.
Note
You are highly encouraged to install these libraries, as they provide large speedups, especially if working with large data sets.
Optional Dependencies¶
Cython: Only necessary to build development version. Version 0.17.1 or higher.
SciPy: miscellaneous statistical functions
PyTables: necessary for HDF5-based storage
matplotlib: for plotting
- statsmodels
- Needed for parts of pandas.stats
Note
Without the optional dependencies, many useful features will not work. Hence, it is highly recommended that you install these. A packaged distribution like the Enthought Python Distribution may be worth considering.
Installing from source¶
Note
Installing from the git repository requires a recent installation of Cython as the cythonized C sources are no longer checked into source control. Released source distributions will contain the built C files. I recommend installing the latest Cython via easy_install -U Cython
The source code is hosted at http://github.com/pydata/pandas, it can be checked out using git and compiled / installed like so:
git clone git://github.com/pydata/pandas.git
cd pandas
python setup.py install
Make sure you have Cython installed when installing from the repository, rather then a tarball or pypi.
On Windows, I suggest installing the MinGW compiler suite following the directions linked to above. Once configured property, run the following on the command line:
python setup.py build --compiler=mingw32
python setup.py install
Note that you will not be able to import pandas if you open an interpreter in the source directory unless you build the C extensions in place:
python setup.py build_ext --inplace
The most recent version of MinGW (any installer dated after 2011-08-03) has removed the ‘-mno-cygwin’ option but Distutils has not yet been updated to reflect that. Thus, you may run into an error like “unrecognized command line option ‘-mno-cygwin’”. Until the bug is fixed in Distutils, you may need to install a slightly older version of MinGW (2011-08-02 installer).
Running the test suite¶
pandas is equipped with an exhaustive set of unit tests covering about 97% of the codebase as of this writing. To run it on your machine to verify that everything is working (and you have all of the dependencies, soft and hard, installed), make sure you have nose and run:
$ nosetests pandas
..........................................................................
.......................S..................................................
..........................................................................
..........................................................................
..........................................................................
..........................................................................
..........................................................................
..........................................................................
..........................................................................
..........................................................................
.................S........................................................
....
----------------------------------------------------------------------
Ran 818 tests in 21.631s
OK (SKIP=2)