Catch cdscan errors without triggering runtime failures about cdms HOT 8 OPEN

@durack1
2 files have the same units time:units = "days since 0301-01-01" ;
The message is very confusing though.

Can you provide me with something better?

Variable 'zg' is duplicated, and is a function of lat or lon:
files zg_Amon_HadGEM2-AO_piControl_r1i1p1_020101-030012.nc, zg_Amon_HadGEM2-AO_piControl_r1i1p1_030101-040012.nc

/cmip5_css02/data/cmip5/output1/NIMR-KMA/HadGEM2-AO/piControl/mon/atmos/Amon/r1i1p1/zg/1/zg_Amon_HadGEM2-AO_piControl_r1i1p1_020101-030012.nc
/cmip5_css02/data/cmip5/output1/NIMR-KMA/HadGEM2-AO/piControl/mon/atmos/Amon/r1i1p1/zg/1/zg_Amon_HadGEM2-AO_piControl_r1i1p1_030101-040012.nc

from cdms.

Similar time overlap of these 2 files. Actually the file timestamps gave a clue...

The first date is 2005-01-16 12 in both file.

/cmip5_css02/scratch/cmip5/output1/NCAR/CCSM4/rcp26/mon/land/Lmon/r1i1p1/v20120416/mrro/mrro_Lmon_CCSM4_rcp26_r1i1p1_200501-210012.nc

and

/cmip5_css02/scratch/cmip5/output1/NCAR/CCSM4/rcp26/mon/land/Lmon/r1i1p1/v20120416/mrro/mrro_Lmon_CCSM4_rcp26_r1i1p1_200601-210012.nc

from cdms.

@dnadeau4 the reason this issue was generated was that when called in a script, the runtime errors (rather than stdout, stderr output) was causing problems. In most cases that I am aware of, the issue is with the input data, so there are problems, but the way that cdscan behaves isn't helpful..

from cdms.

@durack1 you can call cdscan from inside a script

cdms2.cdscan.main(["cdscan","-x","crap.xml","*.nc"])

from cdms.

@doutriaux1 that's exactly what I am doing, but the issue is that runtime failures then trip scripts over.. I was suggesting that as cdscan is a script that is often called outside of an active python/cdms session, it uses the stdout/stderr rather than runtime errors

from cdms.

I tripped on that message
Variable 'zg' is duplicated, and is a function of lat or lon:

It does not tell us that this is a time issue, and I did not understand why zg as duplicated.
Maybe this would be better
Variable 'zg' has duplicated time and is a function of lat or lon:

from cdms.

@durack1 this is a run time error and the program cannot continue. I could raise another kind of exception like "value error". You will still have the back trace, that is how python stops.

from cdms.

@dnadeau4 the issue for me was rather than returning a stderr python returns a runtime error, back trace and trips over scripts that were calling it.. Is it possible to get python "binaries" to behave the same way as a compiled binary, so stdout/stderr?

from cdms.