[Nexus] NXdata for an "10-channel" high resolution detector

[Carlos Pascual]

On Tue 11 June 2013 08:38:08 Wintersberger, Eugen wrote:
> On Tue, 2013-06-11 at 08:32 +0200, Carlos Pascual wrote:
> > Hi Eugen,
> > 
> > are you using Sardana  foracquiring and  writing the data?
> 
> We are not using Sardana yet, though it is already installed on the
> beamline computer. We will move towards Sardana during the long shutdown
> in autumn.
> 
> > If so, it should be possible to create a custom recorder that writes the
> > concatenated data on-the-fly.
> 
> Sure - one can always write custom IO code. I just wanted to know if
> there is an easier solution ;)

Oh, C'mon!
What I mean is that sardana provides an API for writing custom "recorders" 
which is quite simple.  In fact the idea is to have a recorder for each NX 
application definition....

...or at least that *was* the idea, before I saw your presentation (during 
Tango meeting) at how you handle the NX metadata and structure. Then I thought 
that it would be nice to implement the same approach in a generic NX recorder.

> 
> > Carlos
> > 
> > On Mon 10 June 2013 23:59:14 Wintersberger, Eugen wrote:
> > > Thanks guys for the suggestions
> > > 
> > > Currently we are storing the data like at SOLEIL. However, I would
> > > agree with mark that a single array would be much more comfortable for
> > > plotting. The problem with this is that we are writing data in a
> > > record-style manner. This means at every point a new data item is
> > > appended to the data field of the individual detectors. By doing so we
> > > can ensure that every point data is written to disk - even if the scan
> > > breaks accidentally. The only chance we would have is to do the
> > > concatenation afterwards once the scan is finished.
> > > 
> > > Seems that I have to think about which direction we would go ;)
> > > 
> > > regards
> > > 
> > >   Eugen
> > > 
> > > PS: we had now the first beamtime with the new Lambda detector
> > > developed at DESY which will natively write Nexus files. The uses had
> > > no problem with this so far ;)
> > > 
> > > On Mon, 2013-06-10 at 16:52 +0200, Mark Koennecke wrote:
> > > > Eugen,
> > > > 
> > > > On 06/10/2013 04:20 PM, Wintersberger, Eugen wrote:
> > > > > Hi folks
> > > > > 
> > > > >    I have a question concerning NXdata. We are currently starting
> > > > >    with
> > > > > 
> > > > > Nexus at our beamlines here at DESY. P02 - a powder diffraction
> > > > > beamline - will be the first one to use Nexus.
> > > > > They have a kind of 10-channel detector which is in reality a
> > > > > curved detector consisting of 10 individual point detectors each
> > > > > having an analyzer in front of them. The angular offset between
> > > > > the detectors is about 1 degree. We store each individual detector
> > > > > in a separate NXdetector group. The full scan data can be obtained
> > > > > by concatenating the data and polar_angle fields of each of the 10
> > > > > detectors. The question is now: what should go to the NXdata
> > > > > group? Can I have 10 individual pairs of of data/(variable or x??)
> > > > > in a single NXdata group or do I have to create 10 individual
> > > > > NXdata groups?
> > > > > regards
> > > > > 
> > > > >    Eugen
> > > > 
> > > > The plot you want from a  powder diffractometer is polar_angle versus
> > > > counts.
> > > > Thus I would concatenate in the DAQ software and store that in
> > > > NXdata. If you
> > > > have concerns about the concatenation done in the DAQ you can still
> > > > go back to
> > > > the individual detectors. You may also wish to store the concatenated
> > > > data as a
> > > > virtual NXdetector to make the relationship crystal clear.
> > > > 
> > > > Regards,
> > > > 
> > > >             Mark
> > > > > 
> > > > > _______________________________________________
> > > > > NeXus mailing list
> > > > > NeXus at nexusformat.org
> > > > > http://lists.nexusformat.org/mailman/listinfo/nexus
> > > > 
> > > > _______________________________________________
> > > > NeXus mailing list
> > > > NeXus at nexusformat.org
> > > > http://lists.nexusformat.org/mailman/listinfo/nexus