Hi Hella,
We have a test in our benchmark suite that involves a long global
attribute, and it is still working as in previous versions, issuing
the *** NOTE message.
> ferret_v67
NOAA/PMEL TMAP
FERRET v6.7
Linux(gfortran) 2.6.18-238.1.1.el5 - 05/06/11
21-Oct-11 09:24
yes? sp cp longhistory.nc longhistory_copy.nc
yes? let a = 12
yes? save/append/file=longhistory_copy.nc a
LISTing to file longhistory_copy.nc
*** NOTE: CDF attribute history is too long
*** NOTE: no text could be appended
I've attached the file so you can make this test.
So this must be something more to do with the variables and the time
axis. Let's talk off-line from the group and report back to the
Ferret List when we have an answer.
Ansley
On 10/20/2011 1:06 PM, Hella Riede wrote:
Dear
ferreters,
in order to track the history of datasets, I create very long
history lines. When I try to add a variable to such a dataset,
ferret crashes and exits with the message " **ERROR Ferret crash;
signal = 11". It took me some time to figure out that the crash is
due to the long history attribute. I currently use version 6.70.
Up to at least version 6.40, there was a warning issued:
*** NOTE: CDF attribute history is too long
*** NOTE: no text could be appended
This is now seems not to be issued anymore.
After cutting the history line, I get the following error:
"NetCDF: Variable not found (OPeNDAP/netCDF Error code -49)
timep_bnds not found. Attempt to append irregular coordinates to
NetCDF axis which has no bounds attribute. Write data originally
with the /BOUNDS qualifier."
I found some information on that in the release notes to version
6.3. What I do not understand is that I just would like to add a
variable, not additional time steps, to the file.
! I defined a variable
! d=1: test_all_short.nc
! d=2: test.nc
yes? let/d=1 flight = flight[d=2,gt=slon[d=1]@ASN]
! then tried to append it
yes? save/append/file=test_all_short.nc flight[d=1]
The to-be-appended variable flight has the same grid as any
variable in d=1:
yes? sh grid flight
GRID GMH1
name axis # pts start end
normal X
normal Y
normal Z
TIMEP TIME 607 i 28-APR-2006 11:57
28-APR-2006 13:40
yes? sh grid slon
GRID GMH1
name axis # pts start end
normal X
normal Y
normal Z
TIMEP TIME 607 i 28-APR-2006 11:57
28-APR-2006 13:40
In my opinion there is no need for bounds when I really do not
want to extend the file in time.
Any clarification on this is very welcome, thanks in advance!
Best regards,
Hella
|