CDO

hi Lyndon!

There is not operators that does the selection of multiple point on one run. But I would loop over the points, extract each point into a separate file and compute the average of all files at the end. Since these output files should be small, there no big IO overhead.

• selection:

  cdo -remapnn,lon=${lon}_lat=$lat <inputfile> out_${lon}_${lat}

• average:

  cdo  -timmean -cat [ out_* ] all_mean

I used cat which out the values in a temporal order, but since you only the need the average, I think it does not matter

Hi Ralf,

Many thanks for the fast response. I am encountering a segmentation error.

Here's the modified script.

#!/bin/bash
while IFS=, read -r field1 field2 field3 field4 field5 field6 field7
do
    date=$field1
    lat=$field2
    lon=$field3

    latn=$field4
    lats=$field5
    lone=$field6
    lonw=${field7}

echo $lonw

    cdo remapnn,lon=${lon}_lat=${latn} -seldate,${date} vort.2008.850hPa.nc n_out_${date}_${lon}_${latn}
    cdo remapnn,lon=${lon}_lat=${lats} -seldate,${date} vort.2008.850hPa.nc s_out_${date}_${lon}_${lats}
    cdo remapnn,lon=${lone}_lat=${lat} -seldate,${date} vort.2008.850hPa.nc e_out_${date}_${lone}_${lat}
    cdo remapnn,lon=${lonw}_lat=${lat} -seldate,${date} vort.2008.850hPa.nc w_out_${date}_${lonw}_${lat}

    cdo timmean n_out_${date}_${lon}_${latn} s_out_${date}_${lon}_${lats} e_out_${date}_${lone}_${lat} w_out_${date}_${lonw}_${lat} test_${date}

    cdo outputtab,${date},${lat},{$lon},value test_${date} > out_$date.txt
#    dat=$(cdo output test_$date)
#    echo $dat > sample_$date
#    cdo -timmean -cat [ out_* ] all_mean

done < test.dat

I tried this for the first line of the input file

2008-01-19    12.5    127.5    13.5    11.5    128.5    126.5

Here's what I got:

126.5
cdo(1) seldate: Process started
Warning (cdfScanVarAttr): NetCDF: Variable not found - time_bnds
cdo    remapnn: Nearest neighbor weights from lonlat (144x73) to lonlat (1x1) grid
cdo(1) seldate: Processed 10512 values from 1 variable over 20 timesteps.
cdo    remapnn: Processed 10512 values from 1 variable over 1 timestep [0.00s 28MB].
cdo(1) seldate: Process started
Warning (cdfScanVarAttr): NetCDF: Variable not found - time_bnds
cdo    remapnn: Nearest neighbor weights from lonlat (144x73) to lonlat (1x1) grid
cdo(1) seldate: Processed 10512 values from 1 variable over 20 timesteps.
cdo    remapnn: Processed 10512 values from 1 variable over 1 timestep [0.00s 28MB].
cdo(1) seldate: Process started
Warning (cdfScanVarAttr): NetCDF: Variable not found - time_bnds
cdo    remapnn: Nearest neighbor weights from lonlat (144x73) to lonlat (1x1) grid
cdo(1) seldate: Processed 10512 values from 1 variable over 20 timesteps.
cdo    remapnn: Processed 10512 values from 1 variable over 1 timestep [0.00s 27MB].
cdo(1) seldate: Process started

_lat=12.5!apnn (Abort): Open failed on lon=126.5
read_csv.sh: line 30: 2897595 Segmentation fault      (core dumped) cdo remapnn,lon=${lonw}_lat=${lat} -seldate,${date} vort.200850hPa.nc w_out_${date}_${lonw}_${lat}

cdo (Abort): Unprocessed Input, could not process all Operators/Files

cdo    outputtab (Abort): Key 2008-01-19 unsupported!

I tried doing it manually for the first line:

 cdo remapnn,lon=126.5_lat=12.5 -seldate,2008-01-19 vort.2008.850hPa.nc test.nc

And this works! Im not sure what is causing the error.

Also I would like to save the values in a text file:

 cdo outputtab,${date},${lat},{$lon},value test_${date} > out_$date.txt

But it say that the date is unsupported.

Any ideas how to solve this?

-Lyndz

hi Lyndz!

I checked the loop over the uploaded CSV file and got dates, which cannot be processed:

+ IFS=,
+ read -r field1 field2 field3 field4 field5 field6 field7
+ date=21/01/2008
+ lat=12.5
+ lon=120
+ latn=13.5
+ lats=11.5
+ lone=121
+ lonw=$'119\r'
+ echo '|21/01/2008|12.5|120|'
|21/01/2008|12.5|120|
+ cdo remapnn,lon=120_lat=13.5 -seldate,21/01/2008 vort.2008.850hPa.nc n_out_21/01/2008_120_13.5
cdo(1) seldate: Process started

cdo(1) seldate (Abort): Date string >21/01/2008< contains invalid character at position 3!
...

Since you seem to get a different date string from your input I think you might re-upload your csv file. I should at least be able to reproduce your results.

cheers
ralf

Hi Ralf,

the format changed when I uploaded the file. The date column should be YYYY-MM-DD format.

Attached is a text file with the correct format.

Sincerely,