Yes, if you have already a clear data structure it's better to persist
the view as an physical case.
In my case I have some dataset defined in longitude from -180 to 180,
but some users want it from 0 to 360 ... then I create a virtual
dataset, with 2 logical views from the data and one tiled aggregation of
them. This provided to me 2 views served from TDS but from the same
physical dataset without replication.
Regards
Antonio
On 21/12/16 14:25, Vincent THIERION wrote:
Antonio,
Thank you for your answer.
I think my best option is to stich / spatially aggregate individual
files (not virtually) and serve the aggregated image with TDS.
Vincent
Le 21/12/2016 à 11:22, Antonio S. Cofiño a écrit :
Vicent,
On 21/12/16 08:58, Vincent THIERION wrote:
Hello,
By reading different posts on tiled aggregation (i.e. aggregation
along longitude and latitude dimensions), such as :
*
http://www.unidata.ucar.edu/mailing_lists/archives/thredds/2011/msg00270.html
*
http://www.unidata.ucar.edu/mailing_lists/archives/thredds/2013/msg00500.html
better than "labor-intensive" NCML implementation it looks like this
capability is handled by GridDataset and the Netcdf SubsetService
object :
http://www.unidata.ucar.edu/mailing_lists/archives/netcdfgroup/2011/msg00133.html,
only accessible with java library while I work with the python library.
All these posts are quite old and I ask me what the better way is
now ? Maybe stich all the netcdf files in one NetCDF before serve on
TDS ?
The tiled aggregation it's one problem and subsetting is different
one. To aggregate on non-outer dimensions the only existing in
netcdf-java it's tiled one. Never released with full support but it's
there and working. The most common cases are aggregation on
outer-dimensions or non-existing ones.
This aggregation could be made using netcdf-java ncml and serving,
using TDS, as opendap/netcdfsubsetting endpoints.
To subsetting dataset (aggregated on non-aggregated) relies on
tool/client side. Python pyDap or Netcd4 packages are able to access
opendap services to subsetting purposes. Please note that the
Netcdf4 package needs to access to netcdf library compiled with
opendap reading capabilities.
Only specific implementation like x-array python packages has some
multifile aggregation
http://xarray.pydata.org/en/stable/generated/xarray.open_mfdataset.html#xarray.open_mfdataset
NcML aggregation is general specification but only netcdf-java has
full support of it. Some other implemetations exist:
https://github.com/axiom-data-science/pyncml
or the BES implemetation:
http://docs.opendap.org/index.php/BES_-_Modules_-_NcML_Module
But neither of them supports Tiled aggregation.
Regards
Antonio S. Cofiño
Thank you for your useful feedbacks
--
-------------------------------
Vincent THIERION
Ingénieur de recherche - THEIA - CES OSO
vincent.thierion [at] cesbio.cnes.fr
vincent.thierion [at] inra.fr
CESBIO (Bureau 123)
(UMR 5126 CNRS/CNES/UPS/IRD)
Bpi 2801
18 Av. Edouard BELIN
31401 TOULOUSE CEDEX 9
INRA - Toulouse - Dynafor
(UMR 1201 INRA/INP-ENSAT/INP-EI Purpan)
24 chemin de Borde-Rouge
Auzeville BP 52627
31326 CASTANET-TOLOSAN CEDEX
-------------------------------
_______________________________________________
NOTE: All exchanges posted to Unidata maintained email lists are
recorded in the Unidata inquiry tracking system and made publicly
available through the web. Users who post to any of the lists we
maintain are reminded to remove any personal information that they
do not want to be made public.
thredds mailing list
thredds@xxxxxxxxxxxxxxxx
For list information or to unsubscribe, visit:http://www.unidata.ucar.edu/mailing_lists/
_______________________________________________
NOTE: All exchanges posted to Unidata maintained email lists are
recorded in the Unidata inquiry tracking system and made publicly
available through the web. Users who post to any of the lists we
maintain are reminded to remove any personal information that they
do not want to be made public.
thredds mailing list
thredds@xxxxxxxxxxxxxxxx
For list information or to unsubscribe, visit:http://www.unidata.ucar.edu/mailing_lists/
--
-------------------------------
Vincent THIERION
Ingénieur de recherche - THEIA - CES OSO
vincent.thierion [at] cesbio.cnes.fr
vincent.thierion [at] inra.fr
T. +33.(0)5.61.55.85.34
-https://dynafor.toulouse.inra.fr/dynafornet/index.php/fre/Collaborateurs/Permanents/Thierion
CESBIO (Bureau 123)
(UMR 5126 CNRS/CNES/UPS/IRD)
Bpi 2801
18 Av. Edouard BELIN
31401 TOULOUSE CEDEX 9
INRA - Toulouse - Dynafor
(UMR 1201 INRA/INP-ENSAT/INP-EI Purpan)
24 chemin de Borde-Rouge
Auzeville BP 52627
31326 CASTANET-TOLOSAN CEDEX
-------------------------------
_______________________________________________
NOTE: All exchanges posted to Unidata maintained email lists are
recorded in the Unidata inquiry tracking system and made publicly
available through the web. Users who post to any of the lists we
maintain are reminded to remove any personal information that they
do not want to be made public.
thredds mailing list
thredds@xxxxxxxxxxxxxxxx
For list information or to unsubscribe, visit:
http://www.unidata.ucar.edu/mailing_lists/
