Re: Problem with stride when accessing multiple files

From: David Brown <dbrown_at_nyahnyahspammersnyahnyah>
Date: Tue, 26 May 2009 16:03:24 -0600

Hi Ray,

There is a known bug in the current implementation of the code that
strides over
the aggregated dimension of a file list. Even if it appears to work I
would suggest that you do not
trust the results. We have developed a fix and it is due to be
released very soon now
as version 5.1.1. If you would like the new code sooner email Mary
Haley directly.
Be sure to include the results of
uname -a
gcc --version
and if you are using gfortran
gfortran --version


On May 26, 2009, at 3:48 PM, Raymond Arritt wrote:

> Hi folks,
> I've run into an odd problem when processing a long series of data
> using a stride. I've got a bunch of yearly input files that I'm
> treating as a unit, using addfiles and the "cat" option,
> flist = systemfunc ("ls /home/rwarritt/NARCCAP/zg500_NARR_*")
> a1 = addfiles(flist,"r")
> ListSetType (a1, "cat")
> These files have data at 8 times per day (i.e., every 3 hours). I'm
> trying to build a subset over the whole record at one time per day
> using a stride of 8, like this:
> zg500 = a1[:]->zg500(::8,:,:)
> This follows Example 4 at
> except that my input files are 3D instead of 4D as in the example.
> When I do this, the resulting zg500 has all the time samples each day
> as if the stride isn't working. But if I give an explicit lower
> limit, like this,
> zg500 = a1[:]->zg500(0::8,:,:)
> then the stride works as expected and gives zg500 with data only for
> 00 UTC each day.
> I'm curious as to why the stride works only when a lower limit is
> specified. Conversely if I've done something dumb please let me know.
> Thanks
> Ray
> --
> "You miss 100% of the shots you don't take." -- Wayne Gretzky
> _______________________________________________
> ncl-talk mailing list
> List instructions, subscriber options, unsubscribe:

ncl-talk mailing list
List instructions, subscriber options, unsubscribe:
Received on Tue May 26 2009 - 16:03:24 MDT

This archive was generated by hypermail 2.2.0 : Mon Jun 01 2009 - 07:47:27 MDT