The file is corrupted. I ran it against ncvalidator and got the following
errors.
% ncvalidator minc1-no-att.mnc
Error - variable begin offset orders:
var "rootvariable" begin offset (2840) is less than previous variable
"study" end offset (6876)
File "minc1-no-att.mnc" fails to conform with classic CDF file format
specifications
Running it against ncoffsets shows starting file offsets of all variables below.
% ncoffsets minc1-no-att.mnc
netcdf minc1-no-att.mnc {
// file format: CDF-1
file header:
size = 2836 bytes
extent = 6872 bytes
dimensions:
zspace = 10
yspace = 20
xspace = 20
fixed-size variables:
int study:
start file offset = 6872
end file offset = 6876
int rootvariable:
start file offset = 2840
end file offset = 2844
int zspace:
start file offset = 2844
end file offset = 2848
int yspace:
start file offset = 2848
end file offset = 2852
int xspace:
start file offset = 2852
end file offset = 2856
double image-max:
start file offset = 2856
end file offset = 2864
double image-min:
start file offset = 2864
end file offset = 2872
byte image(zspace, yspace, xspace):
start file offset = 2872
end file offset = 6872
}
ncvalidator and ncoffsets are two utility programs available from PnetCDF
package.
It can be compiled independently from PnetCDF with just gcc command.
Wei-keng
> On Jun 22, 2018, at 7:03 PM, Dave Allured - NOAA Affiliate
> <dave.allured@xxxxxxxx> wrote:
>
> Roy,
>
> I think you had a bad download. I got a valid netcdf-3 file,
> minc1-no-att.mnc, 6872 bytes. The full ncdump output is attached. I had to
> add -O and quoting to the curl command as originally shown, and remove the
> extra file name.
>
> I have no clue as to the original issue with scipy.
>
> --Dave
>
>
> On Fri, Jun 22, 2018 at 5:19 PM, Roy Mendelssohn - NOAA Federal
> <roy.mendelssohn@xxxxxxxx> wrote:
> I downloaded the file. the standard "ncdump" from the netcdf library does
> not view the file as a valid netcdf file. Therein lies your problem.
>
> -Roy
>
>
> > On Jun 22, 2018, at 3:27 PM, Ankur Sinha <sanjay.ankur@xxxxxxxxx> wrote:
> >
> > Hello,
> >
> > While building nibabel for Fedora, we noticed a few tests failing when
> > nibabel uses scipy.netcdf[1]. As the nibabel developers say there, the
> > error does not occur with nibabel's bundled copy of netcdf.py, but it is
> > easily reproducible with scipy's version:
> >
> > curl -L
> > https://github.com/nipy/nibabel/blob/master/nibabel/tests/data/minc1-no-att.mnc?raw=true
> > minc1-no-att.mnc
> > python -c 'import scipy.io as sio; sio.netcdf_file("minc1-no-att.mnc")'
> >
> > Would someone please be able to have a look at the scipy issue[2] and
> > provide some hints on where the bug may be? Unfortunately, we're not
> > too well versed with netcdf to diagnose the issue ourselves.
> >
> > [1] https://github.com/nipy/nibabel/issues/579
> > [2] https://github.com/scipy/scipy/issues/8153
> >
> > --
> > Thanks,
> > Regards,
> >
> > Ankur Sinha "FranciscoD"
> >
> > https://fedoraproject.org/wiki/User:Ankursinha
> <minc1-no-att.mnc.txt>_______________________________________________
> 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.
>
>
> netcdfgroup mailing list
> netcdfgroup@xxxxxxxxxxxxxxxx
> For list information or to unsubscribe, visit:
> http://www.unidata.ucar.edu/mailing_lists/
