`pip install pandas` gives UnicodeDecodeError: 'ascii' codec can't decode byte 0xe2 in position 41: ordinal not in range(128)

Question

When performing pip install pandas on a Digital Ocean 512MB droplet, I get the error UnicodeDecodeError: 'ascii' codec can't decode byte 0xe2 in position 41: ordinal not in range(128) .

Any ideas what may have caused it? I'm running Ubuntu 12.04 64bit.

[Full Error]

Solution

It looks like gcc being killed due to insufficient memory (see @Blender's comment) exposed a bug in pip. It mixes bytestrings and Unicode while logging that leads to:

>>> '\n'.join(['bytestring with non-ascii character ☺', u'unicode'])
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
UnicodeDecodeError: 'ascii' codec can't decode byte 0xe2 in position 36: \
   ordinal not in range(128)

If it is reproducible with the latest pip version; you could report the bug.

Comments

[Athena Wisdom]

@Blender just added to the original post

[Blender]

How much RAM do you have on your VPS? GCC is being killed by the OS, which probably means that you ran out of RAM while compiling. Either add a swap partition or just install Pandas via apt-get: sudo apt-get install python-pandas.

[DSM]

Yeah, stuff like this has happened in similar situations to other people.

[TheGrimmScientist]

and to upgrade, pip install --upgrade pip (upgrading fixed this for me)

[Sam]

On ubuntu I could not upgrade pip until I removed the ubuntu package using apt-get remove python-pip and then running the above upgrade code.