Pandas - Revision history

Admin at 07:17, 16 April 2017

2017-04-16T07:17:16Z

← Older revision		Revision as of 07:17, 16 April 2017
Line 219:		Line 219:
	df_csv = pandas.read_csv('dummy.csv')		df_csv = pandas.read_csv('dummy.csv')
	</pre>		</pre>

			{{LinearAlgebraFlag}}
			{{ScientificComputingFlag}}
			{{PythonFlag}}

Admin: /* Installing */

2015-01-15T18:05:49Z

Installing

← Older revision		Revision as of 18:05, 15 January 2015
Line 1:		Line 1:
	=Installing=		=Installing=

	~~For some reason~~, ~~my installation process was a real pain to get working~~, ~~but~~ in ~~the end it all turned out~~ to ~~be these phantom problems~~, ~~that would shift~~, ~~and would act really weird~~, ~~and probably had to do with a crusty Bash <code>$PYTHONPATH</code> variable or something~~.		Installing Pandas can be thorny if you're running on a Mac, mainly because if you download and install your own version of Python, it will conflict with Mac's built-in version of Python. (I recommend leaving Mac's Python version alone.) Mac's version does NOT have pip. This means that if you use pip to install Pandas, it will install it for one version of Python, but not all versions of Python. If you don't run the right Python, Pandas will not be available.

	~~My easy_install and pip were both throwing errors trying to~~ install ~~pandas, due to lacking prerequisites. One had a problem with a too-old~~ version of ~~numpy (not the most recent version~~ that ~~I'd built myself, which~~ is ~~why I think there was some incorrect Python version being sucked in or something). I was able to use [[Python#Virtual_Python:_virtualenv\|Virtualenv]] to debug some of~~ the problems, but basically cleared out my <code>$PYTHONPATH</code>, moved my custom-installed numpy off of my PYTHONPATH, breaking my version of scipy and ipython and matplotlib as well. So, I had to use pip to re-install pip's version of all of these. Installing these ended up not working, however, and when I put my old, by~~-hand installations back on my PYTHONPATH, installing pandas, pytables, and numexpr with pip all went smoothly.~~		When you install your own version of Python, make sure that it is the first python on your path, by typing:

	~~So, in~~ the ~~end, I don't actually know what~~ the right ~~procedure is; I just have a vague sense that there were some problems, that got resolved, by something I did, at some point~~.		<pre>
			which -a python
			</pre>

			This will ensure that the pip on your path corresponds to the right python on your path.

	First, I downloaded and installed easy_install from source.		First, I downloaded and installed easy_install from source.

	Then~~, I blasted my~~ PYTHONPATH:		Then blast your PYTHONPATH (keep things simple):

	<source lang="bash">		<source lang="bash">
Line 26:		Line 30:
	</source>		</source>

	To upgrade:		Or to upgrade:

	<source lang='bash">		<source lang='bash">

Admin at 18:02, 23 July 2013

2013-07-23T18:02:22Z

← Older revision		Revision as of 18:02, 23 July 2013
Line 21:		Line 21:
	$ sudo pip install numpy		$ sudo pip install numpy
	$ sudo pip install numexpr		$ sudo pip install numexpr
			$ sudo pip install cython
	$ sudo pip install tables		$ sudo pip install tables
	$ sudo pip install pandas		$ sudo pip install pandas
			</source>

			To upgrade:

			<source lang='bash">
			$ sudo pip install --upgrade pandas
	</source>		</source>

Admin at 17:56, 23 July 2013

2013-07-23T17:56:09Z

← Older revision		Revision as of 17:56, 23 July 2013
Line 20:		Line 20:
	$ sudo easy_install pip		$ sudo easy_install pip
	$ sudo pip install numpy		$ sudo pip install numpy
	~~$ sudo pip install pandas~~
	~~</source>~~

	~~and installing tables stuff:~~

	~~<source lang="bash">~~
	$ sudo pip install numexpr		$ sudo pip install numexpr
	$ sudo pip install tables		$ sudo pip install tables
			$ sudo pip install pandas
	</source>		</source>

Admin at 02:44, 10 July 2013

2013-07-10T02:44:32Z

← Older revision		Revision as of 02:44, 10 July 2013
Line 1:		Line 1:
	=Installing=		=Installing=

	~~This~~ was a ~~royal~~ pain in the ~~ass~~.		For some reason, my installation process was a real pain to get working, but in the end it all turned out to be these phantom problems, that would shift, and would act really weird, and probably had to do with a crusty Bash <code>$PYTHONPATH</code> variable or something.

	~~I had some problems with my~~ easy_install and pip, ~~probably b/c they were~~ not ~~associated with~~ the ~~right versions of~~ Python. I was able to use [[Python#Virtual_Python:_virtualenv\|Virtualenv]] to debug some of the problems, ~~and~~ basically ~~ended up having to remove~~ my custom-installed ~~version of~~ numpy ~~in favor~~ of ~~the pip-installed version of numpy. This broke~~ my ~~custom-installed~~ version of scipy and ipython as well, so I had to install ~~those using~~ pip. ~~These~~ ended up not working, and I ~~had to go back to~~ my old, by-hand installations~~. So~~, ~~in the end~~, ~~I don't actually know what the right procedure is; I just have a vague sense that there were some problems~~, ~~that got resolved, by something I did, at some point~~.		My easy_install and pip were both throwing errors trying to install pandas, due to lacking prerequisites. One had a problem with a too-old version of numpy (not the most recent version that I'd built myself, which is why I think there was some incorrect Python version being sucked in or something). I was able to use [[Python#Virtual_Python:_virtualenv\|Virtualenv]] to debug some of the problems, but basically cleared out my <code>$PYTHONPATH</code>, moved my custom-installed numpy off of my PYTHONPATH, breaking my version of scipy and ipython and matplotlib as well. So, I had to use pip to re-install pip's version of all of these. Installing these ended up not working, however, and when I put my old, by-hand installations back on my PYTHONPATH, installing pandas, pytables, and numexpr with pip all went smoothly.

	~~Like~~ I ~~said~~, ~~it was a big damn mess~~.		So, in the end, I don't actually know what the right procedure is; I just have a vague sense that there were some problems, that got resolved, by something I did, at some point.

	First, I downloaded and installed easy_install from source.		First, I downloaded and installed easy_install from source.

			Then, I blasted my PYTHONPATH:

			<source lang="bash">
			$ unset PYTHONPATH
			</source>

	Then, I ran the following commands:		Then, I ran the following commands:

	<~~pre~~>		<source lang="bash">
	$ sudo easy_install pip		$ sudo easy_install pip
	$ sudo pip install numpy		$ sudo pip install numpy
	$ sudo pip install pandas		$ sudo pip install pandas
	</~~pre~~>		</source>

	and installing tables stuff:		and installing tables stuff:

	<~~pre~~>		<source lang="bash">
	$ sudo pip install numexpr		$ sudo pip install numexpr
	$ sudo pip install tables		$ sudo pip install tables
	</~~pre~~>		</source>

	=Data=		=Data=

Admin: /* Modifying a Table with Data */

2013-07-09T23:51:43Z

Modifying a Table with Data

@@ Line 188: / Line 188: @@
                          [0.0, 0.0, 0.0, 0.0, 0.0]
                          [0.0, 0.0, 0.0, 0.0, 0.0]
 </pre>

Admin at 23:02, 9 July 2013

2013-07-09T23:02:36Z

Show changes

Admin at 22:10, 9 July 2013

2013-07-09T22:10:22Z

← Older revision		Revision as of 22:10, 9 July 2013
Line 3:		Line 3:
	This was a royal pain in the ass.		This was a royal pain in the ass.

	I had some problems with my easy_install and pip, probably b/c they were not associated with the right versions of Python. I was able to use [[~~Vritualenv~~]] to debug some of the problems, and basically ended up having to remove my custom-installed version of numpy in favor of the pip-installed version of numpy. This broke my custom-installed version of scipy and ipython as well, so I had to install those using pip. These ended up not working, and I had to go back to my old, by-hand installations. So, in the end, I don't actually know what the right procedure is; I just have a vague sense that there were some problems, that got resolved, by something I did, at some point.		I had some problems with my easy_install and pip, probably b/c they were not associated with the right versions of Python. I was able to use [[Python#Virtual_Python:_virtualenv\|Virtualenv]] to debug some of the problems, and basically ended up having to remove my custom-installed version of numpy in favor of the pip-installed version of numpy. This broke my custom-installed version of scipy and ipython as well, so I had to install those using pip. These ended up not working, and I had to go back to my old, by-hand installations. So, in the end, I don't actually know what the right procedure is; I just have a vague sense that there were some problems, that got resolved, by something I did, at some point.

	Like I said, it was a big damn mess.		Like I said, it was a big damn mess.

Admin at 22:09, 9 July 2013

2013-07-09T22:09:05Z

← Older revision		Revision as of 22:09, 9 July 2013
Line 1:		Line 1:
	=Installing=		=Installing=

	This was a pain~~, because~~ I had some ~~things that I hadn't worked out~~ with easy_install and pip. I ~~used~~ [[Vritualenv]] to debug some of the problems, and basically ended up having to remove my custom-installed version of numpy in favor of the pip-installed version of numpy.		This was a royal pain in the ass.

			I had some problems with my easy_install and pip, probably b/c they were not associated with the right versions of Python. I was able to use [[Vritualenv]] to debug some of the problems, and basically ended up having to remove my custom-installed version of numpy in favor of the pip-installed version of numpy. This broke my custom-installed version of scipy and ipython as well, so I had to install those using pip. These ended up not working, and I had to go back to my old, by-hand installations. So, in the end, I don't actually know what the right procedure is; I just have a vague sense that there were some problems, that got resolved, by something I did, at some point.

			Like I said, it was a big damn mess.

	First, I downloaded and installed easy_install from source.		First, I downloaded and installed easy_install from source.

Admin: Created page with "=Installing= This was a pain, because I had some things that I hadn't worked out with easy_install and pip. I used Vritualenv to debug some of the problems, and basically en..."

2013-07-09T21:37:30Z

Created page with "=Installing= This was a pain, because I had some things that I hadn't worked out with easy_install and pip. I used Vritualenv to debug some of the problems, and basically en..."

New page

=Installing=

This was a pain, because I had some things that I hadn't worked out with easy_install and pip. I used [[Vritualenv]] to debug some of the problems, and basically ended up having to remove my custom-installed version of numpy in favor of the pip-installed version of numpy.

First, I downloaded and installed easy_install from source.

Then, I ran the following commands:

<pre>
$ sudo easy_install pip
$ sudo pip install numpy
$ sudo pip install pandas
</pre>

and installing tables stuff:

<pre>
$ sudo pip install numexpr
$ sudo pip install tables
</pre>