[SciPy-user] Fast saving/loading of huge matrices

[Ryan Krauss]

So, I wanted to try out pytables and run the benchmark Francesc
posted, but I ran into an error.  I downloaded
http://www.pytables.org/download/preliminary/tables-2.0b2.win32-py2.4.exe
and installed it, but import tables gives this error:

In [1]: import tables
---------------------------------------------------------------------------
exceptions.ImportError                               Traceback (most recent call
 last)

C:\Documents and Settings\Ryan\<ipython console>

C:\Python24\Lib\site-packages\tables\__init__.py
     49 # Import the user classes from the proper modules
     50 from tables.exceptions import *
---> 51 from tables.file import File, openFile, copyFile
     52 from tables.node import Node
     53 from tables.group import Group

C:\Python24\Lib\site-packages\tables\file.py
    244 # It is necessary to import Table after openFile, because it solves a ci
rcular
    245 # import reference.
--> 246 from tables.table import Table
    247
    248

C:\Python24\Lib\site-packages\tables\table.py
     34 import numpy
     35
---> 36 from tables import tableExtension
     37 from tables.conditions import split_condition
     38 from tables.numexpr.compiler import getType as numexpr_getType

ImportError: cannot import name tableExtension



Is the executable not stand alone?

Ryan

On 4/20/07, Gael Varoquaux <gael.varoquaux at normalesup.org> wrote:
> On Fri, Apr 20, 2007 at 09:30:15AM +0200, Francesc Altet wrote:
> > As I said before, be used to recarrays. If you have reasons for sticking
> > with dictionaries, it is straighforward converting a dict into a
> > recarray.
>
> You are most definitely right that most people (including me) don't use
> recarrays enough. However when importing hdf5 I want to keep part of the
> versatility of hdf5: it is hierarchical and can save much more than
> arrays. Dictionnaries can mirror this possibility quite well in Python.
> This is why I think they are well suited for helper functions to do IO
> with hdf5.
>
> As a side note, we do use this richness of hdf5 in our experiment, to
> store say the time of an experimental run, the temperature of the room...
>
> Gaël
> _______________________________________________
> SciPy-user mailing list
> SciPy-user at scipy.org
> http://projects.scipy.org/mailman/listinfo/scipy-user
>