Path: csiph.com!x330-a1.tempe.blueboxinc.net!usenet.pasdenom.info!aioe.org!news.stack.nl!newsfeed.xs4all.nl!newsfeed5.news.xs4all.nl!xs4all!newsgate.cistron.nl!newsgate.news.xs4all.nl!post.news.xs4all.nl!not-for-mail
To: python-list@python.org
From: Peter Otten <__peter__@web.de>
Subject: Re: package extension problem
Date: Mon, 13 Feb 2012 19:28:05 +0100
Organization: None
References: <19027093.1215.1329153496662.JavaMail.geo-discussion-forums@ynel5> <4F394E17.9090108@inrim.it>
Mime-Version: 1.0
Content-Type: text/plain; charset="ISO-8859-1"
Content-Transfer-Encoding: 7Bit
Precedence: list
Newsgroups: comp.lang.python
Message-ID: <mailman.5769.1329157703.27778.python-list@python.org>
Lines: 58
NNTP-Posting-Host: 2001:888:2000:d::a6
Xref: x330-a1.tempe.blueboxinc.net comp.lang.python:20357

Fabrizio Pollastri wrote:

> Ok. To be more clear, consider the real python package Pandas.
> 
> This package defines a Series class and a DataFrame class.
> The DataFrame is a matrix that can have columns of
> different type.
> 
> If I write
> 
> import pandas as pd
> df = pd.DataFrame({'A':[1,2,3],'B':[4,5,6]})
> 
> a data frame with two cols named A and B is created.
> 
> If I write
> 
> col_A = df['A']
> 
> the returned col_A is an instance of Series.
> 
> Now , let suppose that I want to extend some functionality of pandas
> by adding new methods to both Series and DataFrame classes.
> 
> One way to do this is to redefine this classes in a new package
> (new_pandas) as follow
> 
> import pandas as pd
> 
> class Series(pd.Series):
>       ...
>       add new methods
>       ...
> 
> class DataFrame(pd.DataFrame):
>       ...
>       add new methods
>       ...
> 
> When I use the new package as a pandas substitute and write
> 
> import new_pandas as np
> df = np.DataFrame({'A':[1,2,3],'B':[4,5,6]})
> col_A = df['A']
> 
> col_A is an instance of the original pandas and not of the new pandas,
> losing all the added functionality.

A quick look into the pandas source reveals that the following might work:

# untested
class DataFrame(pd.DataFrame):
    @property
    def _constructor(self):
        return DataFrame # your class
    # your new methods