Re: [Pytables-users] recursive walk question

[Andre' Walker-L. <wal...@gm...>]

Hi Anthony,

On Jun 14, 2012, at 11:30 AM, Anthony Scopatz wrote:

> On Wed, Jun 13, 2012 at 8:23 PM, Andre' Walker-Loud <wal...@gm...> wrote:
> Hi All,
> 
> Still trying to sort out a recursive walk through an hdf5 file using pytables.
> 
> I have an hdf5 file with an unknown depth of groups/nodes.
> 
> I am trying to write a little function to walk down the tree (with user input help) until a data file is found.
> 
> I am hoping there is some function one can use to query whether you have found simply a group/node or an actual numpy array of data.  So I can do something like
> 
> if f.getNode('/',some_path) == "data_array":
>    return f.getNode('/',some_path), True
> else:
>    return f.getNode('/',some_path), False
> 
> where I have some function that if the second returned variable is True, will recognize the file as data, where as if it is False, it will query the user for a further path down the tree.
> 
> 
> I suppose I could set this up with a try: except: but was hoping there is some built in functionality to handle this.
> 
> Yup, I think that you are looking for the File.walkNodes() method.  http://pytables.github.com/usersguide/libref.html#tables.File.walkNodes

I wasn't sure how to use walkNodes in an interactive search.  Here is what I came up with so far (it works on test cases I have given it).  Comments are welcome.

One feature I would like to add to the while loop in the second function is an iterator counting the depth of the search.  I want to compare this to the maximum tree/node/group depth in the file, so if the search goes over (maybe my collaborators used createTable instead of createArray) the while loop won't run forever.

Is there a function to ask the deepest recursion into the hdf5 file?

Cheers,

Andre


def is_array(file,path):
    data = file.getNode(path)
    if str(type(data)) == "<class 'tables.array.Array'>":
        found_array = True
    else:
        found_array = False
        for g in file.getNode(path):
            print g
    return data, found_array

def pytable_walk(file):
    found_data = False
    path = ''
    while found_data == False:
        for g in file.getNode('/',path):
            print g
        path_new = raw_input('which node would you like?\n     ')
        path = path+'/'+path_new
        data,found_data = is_array(file,path)
    return path,data