Page History

...

1. The bufr_filter sequentially processes all bufr messages contained in the input files and applies the rules to each of them. Input messages can be written to the output by using the "write" statement. The write statement can be parameterised so that output is sent to multiple files depending on key values used in the output file name. First we write a rules_file containing the following statement:

   Code Block
   write "../data/split/[bufrHeaderCentre:i]_[dataCategory].bufr[editionNumber]";

   Then we create the BUFR file multitype.bufr from three others:

   Code Block
   > mkdir ../data/split > cat ../data/bufr/syno_1.bufr ../data/bufr/goes_87.bufr ../data/bufr/gosat.bufr > ../data/split/multitype.bufr

   Applying this rules_file to the "../data/split/multitype.bufr" bufr file we obtain several files in the ../data/split directory containing messages split according to their key values

   Code Block
   > bufr_filter rules_file ../data/split/multitype.bufr > ls ../data/split 98_0.bufr3 98_3.bufr4 98_5.bufr3 multitype.bufr

   
   

   
   

2. The bufr header information can be accessed without unpacking the data. This rules_file:

   Code Block
   print "[bufrHeaderCentre] [bufrHeaderSubCentre] [masterTablesVersionNumber] [localTablesVersionNumber] [numberOfSubsets]";

   will result in the following output:

   Code Block
   > bufr_filter rules_file ../data/bufr/syno_multi.bufr 98 0 13 1 1 98 0 13 1 1 98 0 13 1 1

   
   

   
   

3. To print values from the data section the messages have to be unpacked. To do that we need to set key unapack to 1. This rules_file:

   Code Block
   set unpack=1; print "block=[blockNumber] station=[stationNumber] lat=[latitude] lon=[longitude] t2=[airTemperatureAt2M]";

   will print out some data values from the specified SYNOP bufr messages.

   Code Block
   > bufr_filter rules_file ../data/bufr/syno_multi.bufr block=1 station=1 lat=70.93 lon=-8.67 t2=274.5 block=1 station=3 lat=77 lon=15.5 t2=268.4 block=1 station=7 lat=78.92 lon=11.93 t2=268.5

   
   

   
   

4. bufr_filter allows defining new keys with the transient keyword. We will further develop the previous example by creating a new key to combine the block number and the station number into the full WMO station id:

   Code Block
   set unpack=1; transient statid=1000*blockNumber+stationNumber; print "statid=[statid] lat=[latitude] lon=[longitude] t2=[airTemperatureAt2M]";

   The result is:

   Code Block
   > bufr_filter rules_file ../data/bufr/syno_multi.bufr statid=1001 lat=70.93 lon=-8.67 t2=274.5 statid=1003 lat=77 lon=15.5 t2=268.4 statid=1007 lat=78.92 lon=11.93 t2=268.5

   
   

   
   

5. We can use conditional statements in bufr_filter. The syntax is:

   Code Block
   if ( condition ) { block of rules } else { block of rules }

   The condition can be made using ==,!= and joining single block conditions with || and &&. The statement can be any valid statement also another nested condition The rules_file below shows how to filter only SYNOP messages with a specific station id:

   Code Block
   set unpack=1; transient statid=1000*blockNumber+stationNumber; if (dataCategory ==0 && statid == 1003) { write "out.bufr"; }

   
   

   
   

6. The switch statement is an enhanced version of the if statement. Its syntax is the following:

   Code Block
   switch (key1) { case val1: # statements case val2: # statements default: # statements }

   The value of the key given as argument to the switch statement is matched against the values specified in the case statements. If there is a match, then the statements corresponding to the matching case are executed. Otherwise, the default case is executed. The default case is mandatory if the case statements do not cover all the possibilities. The "~" operator can be used to match "anything".

   
   

7. To access the keys' attributes use the -> operator. The example below prints the attributes of key nonCoordinatePressure from a SYNOP bufr message.

   Code Block

   print "nonCoordinatePressure=[nonCoordinatePressure] [nonCoordinatePressure->units]"; print "nonCoordinatePressure->code=[nonCoordinatePressure->code!06d]"; print "nonCoordinatePressure->scale=[nonCoordinatePressure->scale]"; print "nonCoordinatePressure->reference=[nonCoordinatePressure->reference]"; print "nonCoordinatePressure->width=[nonCoordinatePressure->width]"; print "nonCoordinatePressure->percentConfidence=[nonCoordinatePressure->percentConfidence] [nonCoordinatePressure->percentConfidence->units]"; print "nonCoordinatePressure->percentConfidence->code=[nonCoordinatePressure->percentConfidence->code!06d]"; print "nonCoordinatePressure->percentConfidence->scale=[nonCoordinatePressure->percentConfidence->scale]"; print "nonCoordinatePressure->percentConfidence->reference=[nonCoordinatePressure->percentConfidence->reference]"; print "nonCoordinatePressure->percentConfidence->width=[nonCoordinatePressure->percentConfidence->width]";

   The result is:

   Code Block

   > bufr_filter rules_file ../data/bufr/syno_1.bufr nonCoordinatePressure=100910 Pa nonCoordinatePressure->code=010004 nonCoordinatePressure->scale=-1 nonCoordinatePressure->reference=0 nonCoordinatePressure->width=14 nonCoordinatePressure->percentConfidence=74 % nonCoordinatePressure->percentConfidence->code=033007 nonCoordinatePressure->percentConfidence->scale=0 nonCoordinatePressure->percentConfidence->reference=0 nonCoordinatePressure->percentConfidence->width=7

   
   

   
   

8. To access keys by rank (i.e. by their occurrence in the message) use the # operator. The example below prints the value from the 4th occurrence of key pressure from a TEMP bufr message. As a reference, we also print all the pressure values found in the message.

   Code Block
   set unpack=1; print "pressure=[#4#pressure] [#4#pressure->units]"; print "pressure=[pressure]";

   The result is:

   Code Block

   > bufr_filter rules_file ../data/bufr/temp_101.bufr pressure=98500 Pa pressure=102000 101800 100000 98500 96400 92500 92100 89700 88100 86100 85000 84400 79400 79000 78300 77300 71900 70000 69400 65100 61200 53400 50000 43900 40000 39900 37800 31600 30000 27500 25000 21200 21000 20600 20400 20000 19300 18400 17000 16600 15100 15000 14600 14000 13400 13200 12900 11100 10800 10000 8960 7630 7000 6420 6190 5770 5320 5000 3970 3570 3190 3090 3000 2820 2630 2400 2340 2050 2000 1680 1530 1500 1380 1300 1210 31600

   
   

   
   

9. It is possible to access elements by conditions imposed on coordinate descriptors. The example below prints the temperature values on temperature significant levels from a TEMP bufr message. For temperature significant levels the key verticalSoundingSignificance=4 and this is what we use in the condition:

   Code Block
   set unpack=1; print "[/verticalSoundingSignificance=4/airTemperature]";

   The result is:

   Code Block
   > bufr_filter rules_file ../data/bufr/temp_101.bufr 272.1 269.5 268.1 267.9 266.7 266.1 264.9 264.9 260.5 260.9 263.5 263.7 261.7 261.9 259.1 258.9 251.5 243.9 238.3 236.7 221.7 212.7 215.5 215.9 214.1 217.3 218.3 217.3 219.3 218.9 219.5 217.9 218.3 217.5 220.3 219.1 220.1 217.3 216.5 217.7 215.9 217.1 213.5 216.1 214.7 216.1 215.3 216.5 213.9 215.3 215.7 212.7 214.1 216.1 213.7 215.3 214.9

   
   

   
   

10. Another example for accessing keys by condition is to read scatterometer data. File asca_139.bufr contains a single message with 2016 subsets in a compressed form. In this case each subset has exactly the same structure: they store one location with several beams and one backscatter value in each beam. To print the backScatter values for beamIdentifier=2 from all the subsets we can simply define the condition like this:

    Code Block
    set unpack=1; print "/beamIdentifier=2/backscatter=[/beamIdentifier=2/backscatter]";

    The result is:

    Code Block

    > bufr_filter rules_file ../data/bufr/asca_139.bufr /beamIdentifier=2/backscatter=-24.6 -24.78 -24.92 -25.05 -25.04 -24.72 -23.83 -22.57 -21.71 -21.76 -21.81 -20.97 -19.97 -19.01 -17.8 -16.22 -14.67 -13.26 -12.02 -11.01 -9.84 -7.31 -8.76 -10.13 -11.36 -12.58 -13.49 -13.87 -13.77 -13.44 -13.42 -13.58 -13.92 -14.6 -15.36 -16.22 -17.11 -17.98 -18.56 -18.58 -18.49 -18.45 -22.66 -22.99 -23.37 -23.85 -24.27 -24.57 -24.54 -24.17 -23.96 -24.47 -24.53 -23.11 -21.62 -20.27 -18.93 -17.42 -15.78 -14.13 -12.6 -11.35 -10.06 -7.38 -8.57 -9.82 -11.43 -12.88 -13.83 -14.25 -14.21 -14.16 -14.32 -14.44 -14.73 -15.21 -15.94 -17 -17.87 -18.64 and many more values ......

    
    

    
    

11. Accessing keys by condition provides the means to read individual subsets from uncompressed data. File synop_multi_subset.bufr contains a single message with several subsets in an uncompressed form. To access a given subset we can simply use key subsetNumber in the filter condition. The example below shows how to read the blockNumber, stationNumber and airTemperaturefor the first 3 subsets in the message:

    Code Block

    set unpack=1; print "subsetNumber=1 blockNumber=[/subsetNumber=1/blockNumber] stationNumber=[/subsetNumber=1/stationNumber] airTempearture=[/subsetNumber=1/airTemperature]"; print "subsetNumber=2 blockNumber=[/subsetNumber=2/blockNumber] stationNumber=[/subsetNumber=2/stationNumber] airTempearture=[/subsetNumber=2/airTemperature]"; print "subsetNumber=3 blockNumber=[/subsetNumber=3/blockNumber] stationNumber=[/subsetNumber=3/stationNumber] airTempearture=[/subsetNumber=3/airTemperature]";

    The result is:

    Code Block
    > bufr_filter rules_file ../data/bufr/synop_multi_subset.bufr subsetNumber=1 blockNumber=1 stationNumber=27 airTempearture=276.45 subsetNumber=2 blockNumber=1 stationNumber=84 airTempearture=266.55 subsetNumber=3 blockNumber=1 stationNumber=270 airTempearture=275.25

    
    

    
    

...

1. Without options a default list of keys is printed. The default list can be different depending on the type of BUFR message.

   Code Block
   > bufr_ls ../data/bufr/syno_multi.bufr

   The result is:

   Code Block

   ../data/bufr/syno_multi.bufr centre masterTablesVersionNumber localTablesVersionNumber typicalDate typicalTime rdbType rdbSubtype rdbtimeDate rdbtimeTime numberOfSubsets localLatitude localLongitude ecmf 13 1 20090124 120000 1 1 20090124 121435 1 70.93 -8.67 ecmf 13 1 20090124 120000 1 1 20090124 121435 1 77 15.5 ecmf 13 1 20090124 120000 1 1 20090124 121435 1 78.92 11.93 3 of 3 messages in ../data/bufr/syno_multi.bufr 3 of 3 total messages in 1 files

   
   

   
   

2. It is allowed to use wildcards in filenames.

   Code Block
   > bufr_ls ../data/bufr/syno_*.bufr

   
   

   
   

3. To list only a specific set of keys use the -p option.

   Code Block
   > bufr_ls -p totalLength,bufrHeaderCentre,bufrHeaderSubCentre ../data/bufr/syno_multi.bufr

   
   

   
   

4. To list only a subset of messages use the -w (where option). Only the 12 UTC messages are listed with the following line.

   Code Block
   > bufr_ls -w typicalTime="120000" ../data/bufr/syno_*.bufr

   
   

   
   

5. All the non-12 UTC messages are listed as follows:

   Code Block
   > bufr_ls -w typicalTime!="120000" ../data/bufr/syno_*.bufr

   
   

   
   

6. To list only the second message from a BUFR file:

   Code Block
   > bufr_ls -w count=2 ../data/bufr/syno_multi.bufr

   
   

   
   

...

1. Set key bufrHeaderCentre in the header and print its value after the change:

   Code Block
   > bufr_set -v -p bufrHeaderCentre -s bufrHeaderCentre=222 ../data/bufr/syno_1.bufr out.bufr

   
   

   
   

...