Peggy,
We have updated our THREDDS catalogs so that there is more subsetting
at the variable and time levels,
see here,
http://tds.hycom.org/thredds/GLBa0.08/expt_90.9.html
Trying to load the entire GLBa0.08/expt_90.9 model/experiment into
Matlab does not work for a lot of users. The reason for this is there
are currently over 900 days of data in this experiment, and each
experiment contains several variable groupings spread across 5
individual files (e.g., _
2d.nc, _
salt.nc, _
temp.nc, _
uvel.nc,
_
vvel.nc). So when you OPeNDAP request then entire expt_90.9
experiment you are in essence touching/opening 4500 NetCDF files at
once, which can take time or be interrupted for numerous reasons
(catalog updates, new data files appending to the experiment, hundreds
of other users doing the exact same thing, etc.).
So, to help thwart the above issues, it is recommended that you use an
OPeNDAP URL that is "tailored" as close to what you actually want to
query in Matlab. This way the server will only be returning back "what
you need".
We have even setup a "latest" catalog for querying the 10-latest days
in the GLBa0.08/expt_90.9 experiment here (it is updated more
frequently to contain the latest and greatest files),
http://tds.hycom.org/thredds/GLBa0.08/latest.html
btw, alternatively you should check out
http://latest.hycom.org/ as a
failover access method (we don't to aggregations on this
latest.hycom.org yet. all access methods use individual files).
Since this "latest.html" catalog has very few files to query, the
response times will be much quicker (assuming you need access to the
latest 10 days of data).
/mike
> --
> --
> You received this message because you are a member of HYCOM.org
> To ask a question, send an email to
fo...@hycom.org
>
>
>