Re: [Pytables-users] Problem browsing the object tree of an hdf5 file

[Francesc A. <fa...@ca...>]

A Dimarts 31 Maig 2005 14:30, phi...@ho... va escriure:
> Hi list,
>
> I'm browsing an hdf5 file using pytables 1.0 with python 2.3 and fill a
> gtk treeview.
> I use a list composed with dictionaries to store the name and the type
> of node in the hdf5 file.
> Here is the method i use:
<stripped...>
> I didn't find a better way to do it. The problems are:
>     - i use lots of for loop -that's a memory waste-

Not necessarily, but I accept that is more CPU intensive.

>    -  i use three differents functions to get the name of node
> (group._v_name, array.hdf5name, Table.hdf5name). Is it appropriate.

Well, this can be somewhat improved. hdf5name is the real name of the
dataset on disk while name (or _v_name) are the names in PyTables
space (memory). You can use _v_name for all nodes for simplicity.

>    - When i see the result, a table which is in a group is displayed
> also in the root node '/'. I don't know why.

Can you provide a simple example that shows this behaviour?

> Is there a better way to do this?

What about:

for node in h5file:
    components_list.append({'name':node._v_name, 'type':node._v_attrs.CLASS=
})


Cheers,

=2D-=20
>0,0<   Francesc Altet =A0 =A0 http://www.carabos.com/
V   V   C=E1rabos Coop. V. =A0=A0Enjoy Data
 "-"