How can I install a specific build version?

nov...@gmail.com

unread,

Feb 12, 2014, 4:14:59 AM2/12/14

to co...@continuum.io

There is both a np17 and np18 version of pandas 0.13.0 in the repo, but when I try to update pandas the np18 version is automatically selected.

Is there some way to tell conda to install a specific build version so that I don't have also upgrade numpy?

Thanks,

Dave

C:\dev\code>conda search pandas

pandas 0.9.0 np16py27_0 defaults

0.9.1 np16py27_0 defaults

0.10.0 np16py27_0 defaults

0.10.1 np16py26_0 defaults

0.10.1 np16py27_0 defaults

0.10.1 np17py26_0 defaults

0.10.1 np17py27_0 defaults

0.10.1 np17py33_0 defaults

0.11.0 np17py26_1 defaults

0.11.0 np17py27_1 defaults

0.11.0 np17py33_1 defaults

0.12.0 np17py26_0 defaults

0.12.0 np17py27_0 defaults

0.12.0 np17py33_0 defaults

0.12.0 np17py26_1 defaults

* 0.12.0 np17py27_1 defaults

0.12.0 np17py33_1 defaults

0.13.0 np17py26_0 defaults

0.13.0 np17py27_0 defaults <================

0.13.0 np17py33_0 defaults

0.13.0 np18py26_0 defaults

0.13.0 np18py27_0 defaults

0.13.0 np18py33_0 defaults

0.13.1 np18py26_0 defaults

0.13.1 np18py27_0 defaults

0.13.1 np18py33_0 defaults

C:\dev\code>conda install pandas==0.13.0

Package plan for installation in environment C:\dev\bin\Anaconda:

The following packages will be downloaded:

package | build

---------------------------|-----------------

numpy-1.8.0 | py27_0 23.8 MB

pandas-0.13.0 | np18py27_0 3.3 MB

scipy-0.13.3 | np18py27_0 55.9 MB

------------------------------------------------------------

Total: 83.0 MB

The following packages will be UN-linked:

package | build

---------------------------|-----------------

numpy-1.7.1 | py27_3

pandas-0.12.0 | np17py27_1

scipy-0.13.2 | np17py27_0

The following packages will be linked:

package | build

---------------------------|-----------------

numpy-1.8.0 | py27_0 hard-link

pandas-0.13.0 | np18py27_0 hard-link

scipy-0.13.3 | np18py27_0 hard-link

Proceed ([y]/n)? n

C:\dev\code>python

Python 2.7.5 |Anaconda 1.8.0 (64-bit)| (default, Jul 1 2013, 12:37:52) [MSC v.1500 64 bit (AMD64)] on win32

Type "help", "copyright", "credits" or "license" for more information.

nov...@gmail.com

unread,

Feb 12, 2014, 4:22:16 AM2/12/14

to co...@continuum.io, nov...@gmail.com

On Wednesday, 12 February 2014 09:14:59 UTC, nov...@gmail.com wrote:

There is both a np17 and np18 version of pandas 0.13.0 in the repo, but when I try to update pandas the np18 version is automatically selected.

Is there some way to tell conda to install a specific build version so that I don't have also upgrade numpy?

Thanks,
Dave

To answer my own question, I can download the bz2 from http://repo.continuum.io/pkgs/free/win-64/ and then install by just giving the filename.

Is there an easier way that this? I think this functionality would be nice to have....

Thanks,

Dave

Ilan Schnell

unread,

Feb 12, 2014, 11:31:21 AM2/12/14

to Dave Hirschfeld, conda

You can use:

> conda install pandas=0.13.0

note the single '=' sign, or to specify a special build:

> conda install pandas=0.13.0=np17py27_0

- Ilan

--
You received this message because you are subscribed to the Google Groups "conda - Public" group.
To unsubscribe from this group and stop receiving emails from it, send an email to conda+un...@continuum.io.
To post to this group, send email to co...@continuum.io.
Visit this group at http://groups.google.com/a/continuum.io/group/conda/.

Aaron Meurer

unread,

Feb 12, 2014, 11:32:23 AM2/12/14

to Ilan Schnell, Dave Hirschfeld, conda

You can also set numpy=1.7 to require it to use the np17 version.

Aaron Meurer

Anthony Scopatz

unread,

Apr 29, 2015, 1:44:38 PM4/29/15

to co...@continuum.io, nov...@gmail.com

Hi Ilan,

Is this documented anywhere? I expected it at http://conda.pydata.org/docs/spec.html but couldn't find it...

Be Well
Anthony

Ilan Schnell

unread,

Apr 29, 2015, 2:42:04 PM4/29/15

to Anthony Scopatz, conda, Dave Hirschfeld

Hi Anthony,

the package specification, e.g. "python=2.6" is explained here:
http://conda.pydata.org/docs/intro.html

- Ilan

Aaron Meurer

unread,

Apr 29, 2015, 3:08:35 PM4/29/15

to Anthony Scopatz, conda, Dave Hirschfeld

On Wed, Apr 29, 2015 at 12:44 PM, Anthony Scopatz <sco...@gmail.com> wrote:

> Hi Ilan,
>
> Is this documented anywhere? I expected it at
> http://conda.pydata.org/docs/spec.html but couldn't find it...

That document describes the technical details of the internals of a
conda package. The ways the command line arguments work differ in many
ways, because the command line syntax aims for ease of declaration on
the command line and the package spec aims for explicitness and
clarity.

Aaron Meurer

Anthony Scopatz

unread,

Apr 29, 2015, 3:46:40 PM4/29/15

to Aaron Meurer, conda, Dave Hirschfeld

Thanks Ilan, Aaron,

I see now that the intro doc say that you can substitute '=' for '-' but there are no examples where the double '=' comes up. Thanks though.