Application Development in Python

Dave Reed drlinux at columbus.rr.com
Wed Jul 9 17:17:18 CEST 2003


On Wednesday 09 July 2003 10:37, John Hunter wrote:
> >>>>> "Dave" == Dave Reed <drlinux at columbus.rr.com> writes:
> 
>     Dave> Also note that using the database for manipulating the data
>     Dave> makes the speed very acceptable. I'm almost certain that
>     Dave> manipulating the data in Python would be way too slow to
>     Dave> make it usable (e.g., generating a balance sheet that looks
>     Dave> through all the transactions and groups them by account in
>     Dave> Python would be much slower than using sql commands to do
>     Dave> that).
> 
> I'm not arguing that the dbase isn't the way to go, but if you wanted
> to do the manipulations in python, the Numeric package certainly
> provides the speed you need to manipulate large quantities of data
> rapidly.  I sometimes use a C extension to put data from an SQL
> database directly into Numeric arrays and then manipulate the data in
> python.
> 
> John Hunter

Yes, Numeric is great for numeric data. The data I have is a mix of
numeric and text data (for example, patient name, frame description,
etc.). The other benefit of using the database is it provides
automatic persistence (permanent storage) and atomic transaction
processing. And of course sql statements allow you to easily get the
data you want using joins, group by, and order by.

Dave







More information about the Python-list mailing list