[BioPython] Problems with NCBIXML.py

[Bruno Santos]

Peter Said:
>XML files are big.  Lots of query sequences will also make things 
>bigger.  And the default expectation threshold will also give lots of 
>results - setting this to something harsher will help by giving less 
>matches.
>
>Unless you want to keep the XML file for other analysis, it might be 
>simpler to parse the output from blast directly with Biopython - 
>avoiding having the large XML file on disk.
>
>Keeping the XML intermediate file can be a good idea when working on 
>smaller datasets, where you want to tweak your analysis (without 
>re-running blast each time).

But if even I don't want to save the results to an XML I still have to do
the <blast_results = result_handle.read() #Catch the results> step right?
And my problem is in this step not in writing to the file. 
Or I can use the result_handle directly, because I was reading the biopython
documentation but it's not very clear.