[SciPy-user] HDF5 vs FITS (was: Fast saving/loading of hugematrices)

Perry Greenfield perry@stsci....
Fri Apr 20 20:54:14 CDT 2007

James Turner wrote:

> Thanks for the interesting summary. My question was a bit vague just
> because I was asking more out of curiosity than for a specific
> application.
> Actually, in the back of my mind I was wondering how SciPy and STScI
> Python might end up playing together... I think SciPy will prove
> rather useful and it obviously helps to talk the same language as
> other users/developers (so thumbs up for NumPy :-) ). Maybe if there
> were a single interface to image/table-type formats, it would allow
> more interaction -- but I imagine such a thing might be impractical
> to implement. Don't worry too much -- I'm just thinking out loud!
Well, the next release of stsci_python will be based on numpy, and that
should be out in early summer.

As far as tables go, certain aspects of FITS table semantics make
compatibility difficult (scaled columns, Null values, etc), though we do
plan to make raw numpy record arrays views of the same data buffer
available. I don't know if that's what you are referring to here or not.


More information about the SciPy-user mailing list