4A2-GRIB API库练习2.pdfVIP

Practical 2: GRIB API indexing cd $SCRATCH/practicals/exercise2 cd F90 # or ‘cd C’ or ‘cd Python’ ls Makefile eps grib_api_index.f90 solution  The file ‘eps’ contains EPS fields for all (50) (perturbed) ensemble members for some (4) parameters, e.g. see output of ‘grib_ls eps’.  The objective of this exercise is to write a Fortran, C code or a Python script using the GRIB API and the indexed access method to compute the ensemble mean for one parameter.  The file grib_api_index.f90 (.c or .py) contains a skeleton of the code or script. Please complete it. If you are confident, you can start from Slide 1 ‘scratch’. com-grib 2013 - grib indexing - grib coding ©ecmwf Slide 1 Practical 2: Main program program index 2 index tables will be use grib_api needed implicit none integer :: iret integer,dimension(:),allocatable :: paramId,number integer :: paramIdSize,numberSize two tables are defined integer :: i,j for the data values. integer :: idx,igrib,count,numberOfValues real (KIND=8),dimension(:), allocatable :: values, result Compile/link with: Fortran: pgf90 –o grib_api_index grib_api_index.f90 $GRIB_API_INCLUDE $GRIB_API_LIB C: gcc -o grib_api_index grib_api_index.c $GRIB_API_INCLUDE Slide 2 $GRIB_API_LIB -lm or use the Makefile. com-grib 2013 - grib indexing - grib coding ©ecmwf Slide 2 Practical 2: GRIB API indexing Run the resulting code with: ./grib_api_index # for Fortran or C python grib_api_index.py # for Python Now change the link for the input file ‘eps’ to the grib2 file (also available from ~trx