| Re: HDF File Question [message #55180] |
Thu, 02 August 2007 02:26 |
James Kuyper
Messages: 425
Registered: March 2000
|
Senior Member |
|
|
David Fanning wrote:
> Folks,
>
> Is it possible to both read from and write to the *same*
> HDF file? In other words, could I read a scientific
> data set, make some changes to it, and save it back
> in the same file? If so, would I have to write it
> back as a different SD, or could I just replace the
> one I read.
Yes, it is perfectly possible. That's what the /RDWR option of
HDF_SD_Start() is for. I've used this feature frequently.
However, keep in mind that the HDF format doesn't allow a very
sophisticated mangagement of file space. What I'm about to say doesn't
apply to SDSs, but it applies to most of the other things you can
write to an HDF file, such as file or SDS attributes. If you write a
new value for an existing attribute, the current value will remain in
the file, but is simply marked as no longer in use. The new value is
added to the file, rather than overwriting the first one. This is
invisible if you locate the attribute by name; you always locate the
last-written value. However, the earlier values can be retrieved by
using low-level HDF routines to walk throught the file one compenent
at a time. This is an understanding that I've picked up somewhere
along the line of more than a decade of working with HDF files - I
can't say that I've ever put this to a direct test. However, I have
watched the file size get steadily bigger when the same file attribute
gets re-written multiple times.
|
|
|
|
comp.lang.idl-pvwave archive
Members
Search
Help
Login
Home
