[netcdfgroup] simple commandline query of data-variable statistics to stdout?

• To: netcdfgroup@xxxxxxxxxxxxxxxx,
• Subject: [netcdfgroup] simple commandline query of data-variable statistics to stdout?
• From: Tom Roche <Tom_Roche@xxxxxxxxx>,
• Date: Thu, 06 Sep 2012 19:03:17 -0400

summary: Suppose you have a netCDF file in the cwd named 'foo.nc'
which has a datavar named 'bar' which contains numeric data ranging
from +1.23e-4 to +5.67e8. I'm looking for executable(s) that supports
use roughly similar to the following commandline input and output:

$ ncmax ./foo.nc bar
5.67e8

details:

I'd like to be able to simply, quickly, and automatically test netCDF
data, e.g. to ensure manipulations I'm making don't make unintentional
changes. One obvious way to do this is to check simple statistics,
like the range (max and min) or central tendency (mean or median), of
a datavar, to ensure it changes (or not) as expected.

I don't particularly care about the technology/API used for the task,
but thought this would be easy to do with NCO, so I tried that first.
After 30 min I still couldn't figure out how, so I punted to google,
where I found

http://sourceforge.net/projects/nco/forums/forum/9830/topic/4008865

which suggests this should be easy to do ... but doesn't say how.
Taking the bait, I examined ncks and ncap2, but, after another 30 min,
was still unenlightened: I can see relatively easily how to create an
output netCDF file, some datavar of which contains the desired value,
but again, that's not what I (and the asker of the above) want--we
just want good old-fashioned stdout.

So I wrote an R script which demonstrates (albeit crudely) something
like the desired functionality. If you have R and Rscript (with R
packages ncdf4 and fields), you should be able to

1 clone https://github.com/TomRoche/GEIA_to_NetCDF to a local folder

2 download

https://github.com/downloads/TomRoche/GEIA_to_netCDF/N2OOC90Y.1A

  to the local folder

3 in that folder, run ./GEIA.to.netCDF.r to create ./GEIA_N2O_oceanic.nc

4 in that folder, run

$ Rscript ./netCDF.stats.to.stdout.r 'netcdf.fp="./GEIA_N2O_oceanic.nc"' 
'var.name="emi_n2o"'
> For ./GEIA_N2O_oceanic.nc var=emi_n2o
>       cells=64800
>       obs=36143
>       min=5.96e-08
>       max=1.17e+03
>       mean=99.5
>       med=67.7

But I strongly suspect there's an easier way--i.e., not involving
writing R--and that someone has scratched this itch before. Am I
missing something?

TIA, Tom Roche <Tom_Roche@xxxxxxxxx>

• Follow-Ups:
  • Re: [netcdfgroup] simple commandline query of data-variable statistics to stdout?
    • From: Lynnes, Christopher S. (GSFC-6102)
  • [netcdfgroup] If a compressed dataset is on a THREDDS server, where is the uncompression done?
    • From: Ansley Manke