0.10.0 PyTables Support

[Aquil H. Abdullah]

Hello All,

I am currently using Pandas for performing investigations on data stored in PyTables. The data is stored in tables by groups. I've looked at the documentation as well as the source code:

https://github.com/wesm/pandas/blob/master/pandas/io/pytables.py

but it still isn't clear to me if or how I can access a table in my HDF5/PyTable file.

Do I need to convert my existing tables to DataFrames and store those in my HDF5 file?

To be clear my tables are mixed type.

I can open the file with:

In [1]: import pandas as pd

In [2]: store = pd.HDFStore('test.h5')

and I can see what nodes are available with:

In [27]: store.root

Out[27]: 

/ (RootGroup) 'Market Data'

  children := ['EQUITY' (Group)]

but accessing the group or the tables gives me PyTable objects. I will continue to grok the IO How-To as well as the pytables.py file, but I guess that my first question is wether or not I need to store my tables as pandas_types?

-- 

Aquil H. Abdullah

"I never think of the future. It comes soon enough" - Albert Einstein

[Jeff Reback]

there has been discussion on this issue:

https://github.com/pydata/pandas/issues/1985

 

but I think you just want to get back the frames, possibily via a query,

 

so store.select('EQUITY') should work

store.keys() will give you all of the groups

 

you really don't need to access the PyTables groups (unless I am missing something)

 

hth,

 

Jeff