[Numpy-discussion] mysql -> record array

Travis Oliphant oliphant.travis at ieee.org
Sat Nov 18 10:06:09 CST 2006


Erin Sheldon wrote:
> Hi Travis -
>
> That is an impressive speed increase.  Why is w/o dtype taking
> so much longer?  Is this just from determining elements sizes and
> counts?
>   
If you don't specify the data-type, then the auto-discovery algorithm 
looks at each element in the nested sequence to determine what the 
data-type should be.  This takes a while.  

-Travis




More information about the Numpy-discussion mailing list