Re: combined_datasets questions

On Tue, Dec 16, 2003 at 10:38 -0500, David P. Huenemoerder wrote:
> 
> Hi John - 
> 
> I'm having some confusion or problems w/ combine_datasets and 
> get_combined.  I loaded the HEG and MEG -1 and +1, arfs, and rmfs, and
> flux_corr'd.  I matched_dataset_grids, and combined_datasets.
> But when I
> 
>   f = get_combined(1, &get_data_flux);
> 
> the result is the sum of the four fluxes, not the mean.


Yes, get_combined merely returns the weighted sum of the
datasets in the specified group, propagating uncertainties by
quadrature sum.

The intent was that weighting could support use of this
function with flux-corrected data.  I did not anticipate using
both flux-corrected and counts simultaneously.

If I add functions to support retrieving and modifying the
weights, it would be easier to use this function with counts
and flux-corrected data simultaneously.  For example, you might
do something like:

  old_weights = get_combination_weights (1);
  set_combination_weights (1, new_weights);
  
  f = get_combined (1, &get_data_flux);
  
  set_combination_weights (1, old_weights);

for a suitable definition of 'new_weights'.


> 
> Also,
> 
>   a = get_combined(1, &get_arf);
> 
> returns something, but it isn't a single, nor summed arf, as far as I
> can tell.  get_combined isn't documented as retrieving an arf, but
> there was no error.
> 

The intent was that the function reference should refer to one
of 
   get_data_counts, get_data_flux
   get_model_counts, get_model_flux,
but the only real requirement is that the function must return
the appropriate structure. get_arf works, but the result may or
may not be appropriate.

I could add code to prevent the user from supplying any
function other than the 4 I mentioned, but that might prevent
some legitimate uses of this function.  I'll try to clarify
this in the documentation.

Thanks,
-John
----
You received this message because you are
subscribed to the isis-users list.
To unsubscribe, send a message to
isis-users-request_at_email.domain.hiddenwith the first line of the message as:
unsubscribe