[Met_help] vsdb tool output

[John Halley Gotway]

Ruifang,

I apologize for the delay in getting back to you.

I think I understand what's going on here.  It's just a bit of confusion as to what exactly the output the VSDB-Analysis tool is.  Let me explain...

The VSDB-Analysis tool really performs 2 steps.  First, it's reads through all of the VSDB lines you pass to it, and it FILTERS out the ones that you'd like to retain by applying all of the filtering
parameters you set in the configuration file or on the command line.  Second, once it has that filtered set of VSDB lines, it runs the jobs that you define over that data.

Now, the two jobs you're running are both of type "vsdb_job_summary".  This job requires you to select two things:
(1) a line type of interest: CNT/0.050 in your case
(2) a column of data from that  line (using the "-column" option) to summarize: in your case, RMSE for job 1 and MBIAS for job 2
The job simply computes summary information for the column of data you choose.  That summary information consists of the mean, standard deviation, min, max, and several percentile values for that
column of data.

So the output of the VSDB-Analysis tool is really that line of summary information for the job.  By default, it will get printed to the screen, but you can use the "-out" command line option to write
it to a file instead.  Also, you can dump the lines of VSDB data that were used in this job using the "-dump_row" option.  The files created through dump_row simply contain the VSDB lines that went
into the job.  We provide the "dump_row" option so that users can verify that the VSDB-Analysis tool is filtering the data and computing the summary statistics the way they expect.

I believe that you're confusing the "dump_row" output with the real job output.  It really does make sense that the files you list in your message would be of exactly the same length.  For example,
you used the exact same VSDB filtering settings when generating the files "cnt_point_stat_SPFH_RMSE.vsdb" and "cnt_point_stat_SPFH_BIAS.vsdb".  The only difference is that you're looking at two
different column from the same set of VSDB lines.  So the output of "dump_row" should be the same, but the summary information that gets computed and printed to the screen should be different.

Hope that helps explain things.  Feel free to write with more questions about it.

Thanks,
John Halley Gotway
johnhg at ucar.edu

Ruifang Li wrote:
> Hi,
> I tried METv 1.1 vsdb-analysis tool to summerize output from point-stat. I
> defined column as RMSE AND MBIAS in config file and get the same output for
> different column. Is there something wrong in my config file?
> 
> Here is my config file for vsdb:
> model[] = [];
> fcst_lead[] = ["12","18","24","30","36","42","48","54","60"];
> valid_beg = "20070815_000000";
> valid_end = "20070817_000000";
> init_beg = "";
> init_end = "";
> init_hour[] = [];
> obtype[] = ["ADPUPA", "ADPSFC", "SFCSHP","VADWND","PROFLR","AIRCFT"];
> vx_mask[] = ["FULL"];
> line_type[] = [];
> var[] = [];
> level[] = [];
> interp_mthd[] = [];
> interp_pnts[] = [];
> joblist[] = [
> "-job vsdb_job_summary  -var $VAR  -line_type CNT/0.050  -dump_row
> $OUTDIR/vsdb_analysis/cnt_point_stat_${VAR}_RMSE.vsdb -column RMSE",
> "-job vsdb_job_summary  -var $VAR  -line_type CNT/0.050  -dump_row
> $OUTDIR/vsdb_analysis/cnt_point_stat_${VAR}_BIAS.vsdb -column MBIAS"
>             ];
> alpha = 0.05;
> boot_interval = 1;
> boot_rep_prop = 1.0;
> n_boot_rep = 1000;
> boot_rng = "mt19937";
> boot_seed = "";
> rank_corr_flag = 1;
> tmp_dir = "/tmp";
> version = "V1.1";
> 
> 
> 
> Here is the output which shows that RMSE and MBIAS file has the same size
> for tmp,wind, ugrd and vgrd. I differed two files and they are same.
> 
> -rw-r--r--    1 lir      ncar         151130 Feb 20 11:59
> cnt_point_stat_SPFH_RMSE.vsdb
> -rw-r--r--    1 lir      ncar         151130 Feb 20 11:59
> cnt_point_stat_SPFH_BIAS.vsdb
> -rw-r--r--    1 lir      ncar         364490 Feb 20 11:59
> cnt_point_stat_UGRD_RMSE.vsdb
> -rw-r--r--    1 lir      ncar         364490 Feb 20 11:59
> cnt_point_stat_UGRD_BIAS.vsdb
> -rw-r--r--    1 lir      ncar         364490 Feb 20 11:59
> cnt_point_stat_VGRD_RMSE.vsdb
> -rw-r--r--    1 lir      ncar         364490 Feb 20 11:59
> cnt_point_stat_VGRD_BIAS.vsdb
> -rw-r--r--    1 lir      ncar         179070 Feb 20 12:00
> cnt_point_stat_TMP_RMSE.vsdb
> -rw-r--r--    1 lir      ncar         179070 Feb 20 12:00
> cnt_point_stat_TMP_BIAS.vsdb
> 
> 
> Thanks for the help,
> Ruifang
> 
> Mesoscale & Microscale Meteorology Division
> Phone: 303-497-8938
> Office:  FL3-3085
> Email:  lir at ucar.edu
> 
> 
> 
> ------------------------------------------------------------------------
> 
> _______________________________________________
> Met_help mailing list
> Met_help at mailman.ucar.edu
> http://mailman.ucar.edu/mailman/listinfo/met_help