Re: Strange problem using FORTRAN code in NCL

From: Wei Huang <huangwei_at_nyahnyahspammersnyahnyah>
Date: Mon, 5 Oct 2009 08:14:28 -0600

You may try to use NCL's string function to read this type data.
Below is a sample script.

Wei Huang
huangwei_at_ucar.edu
VETS/CISL
National Center for Atmospheric Research
P.O. Box 3000 (1850 Table Mesa Dr.)
Boulder, CO 80307-3000 USA
(303) 497-8924

-----------------------

  load "$NCARG_ROOT/lib/ncarg/nclscripts/csm/gsn_code.ncl"
  load "$NCARG_ROOT/lib/ncarg/nclscripts/csm/gsn_csm.ncl"
;load "$NCARG_ROOT/lib/ncarg/nclscripts/csm/contributed.ncl"

  fname = "./L3_aersl_omi_20090101.txt"

;read all the strings
  strs = asciiread(fname, -1, "string")

;get lat/lon info
  lon_str = str_get_field(strs(1), 2, ":")
  lat_str = str_get_field(strs(2), 2, ":")

  nlon = tointeger(str_get_field(lon_str, 1, " "))
  nlat = tointeger(str_get_field(lat_str, 1, " "))

  print(nlon)
  print(nlat)

;define data size
  dat = new((/nlat, nlon/), integer)
  dat@_FillValue = 999

  lon = new(nlon, float)
  lat = new(nlat, float)
  lon_at_long_name = "Longitude"
  lat_at_long_name = "Latitude"
  lon_at_units = "degree_east"
  lat_at_units = "degree_north"

;define an array of new string, and strip of the header lines, and
leading space.
  nlines = dimsizes(strs(3:))
  new_strs = new(nlines, string)

  do n = 0, nlines - 1
     new_strs(n) = str_get_cols(strs(n+3), 1, strlen(strs(n+3)))
  end do

;put one latitude data into one string
  full_lat_strs = new(nlat, string)
  do n = 0, nlat - 1
     lat(n) = -89.5 + n
     ns = n*15
     ne = ns + 14
     full_lat_strs(n) = str_concat(new_strs(ns:ne))
  end do

;print(full_lat_strs)

;get all the data.
  do n = 0, nlon - 1
     lon(n) = -179.5 + n
     ns = 3*n
     ne = ns + 2
     dat(:, n) = tointeger(str_get_cols(full_lat_strs(:), ns, ne))
  end do

  printVarSummary(dat)

---------------------

On Oct 5, 2009, at 7:33 AM, Mary Haley wrote:

> You are defining "ai" to be integer in your Fortran program, but in
> the NCL code, you are defining "var" to be a float, and trying to
> pass this
> to your Fortran routine as the "ai" array.
>
> Try making "var" an integer:
>
> var = new((/180,360/),int)
>
> --Mary
>
> On Oct 3, 2009, at 1:45 PM, eddycarl wrote:
>
>> Sorry, I have found the problem. It is due to the "stop" statement
>> in the fortran code.
>> But, I still have problem reading the correct data value. Now, I
>> can get correct value of "lat" and "lon". However, the value of the
>> 2d array "var", which is printed by NCL, is totally wrong (while,
>> the value printed out by fortran code is correct!). So, I doubt
>> whether NCL has bugs in tranfering 2d array from FORTRAN to NCL?
>> Thanks for your attention!
>>
>> >Send ncl-talk mailing list submissions to > ncl-talk_at_ucar.edu >
>> >To subscribe or unsubscribe via the World Wide Web, visit > http://mailman.ucar.edu/mailman/listinfo/ncl-talk
>> >or, via email, send a message with subject or body 'help' to > ncl-talk-request_at_ucar.edu
>> > >You can reach the person managing the list at > ncl-talk-owner_at_ucar.edu
>> > >When replying, please edit your Subject line so it is more
>> specific >than "Re: Contents of ncl-talk digest..." > > >Today's
>> Topics: > > 1. Re: intel-mac binary 32 or 64 bit? (Jamie Scott)
>> > 2. Re: intel-mac binary 32 or 64 bit? (Wei Huang) > 3.
>> "Height" axis label on left in gsn_csm_pres_hgt > (Oswald
>> Jason Lobo) > 4. Strange problem using FORTRAN code in NCL
>> (eddycarl) > >
>> >
>> ---------------------------------------------------------------------- >
>> >Message: 1 >Date: Fri, 02 Oct 2009 15:39:22 -0600 >From: Jamie
>> Scott <James.D.Scott_at_noaa.gov> >Subject: Re: intel-mac
>> binary 32 or 64 bit? >To: Wei Huang <huangwei_at_ucar.edu>, ncl-talk_at_ucar.edu
>> >Message-ID: <A4573765-54A2-461C-9A26-1E70B34F2885_at_noaa.gov>
>> >Content-Type: text/plain; charset="us-ascii" > >Doesn't seem to be
>> any limit to memory. Also "top" indicates I have >3.3GB
>> available. > >> cputime unlimited >> filesize unlimited >>
>> datasize 6144 kbytes >> stacksize 8192 kbytes >>
>> coredumpsize 0 kbytes >> memoryuse unlimited >> descriptors 256
>> >> memorylocked unlimited >> maxproc 266 > > >-Jamie >On Oct
>> 2, 2009, at 3:32 PM, Wei Huang wrote: > >> Jamie, >> >> Your Mac
>> has 8G memory, does not mean your NCL can use all of them. >> You
>> may have other program running, such as your dock, web browser,
>> >> finder, etc. >> For example, I have firefox running, which
>> takes about 150M. >> >> Also, there may be per process limit on
>> your mac, and you may >> use "limit" command to see if there is a
>> limit on each process. >> >> Thanks, >> >> Wei Huang >> huangwei_at_ucar.edu
>> >> VETS/CISL >> National Center for Atmospheric Research >> P.O.
>> Box 3000 (1850 Table Mesa Dr.) >> Boulder, CO 80307-3000 USA >>
>> (303) 497-8924 >> >> >> >> >> >> On Oct 2, 2009, at 3:10 PM, Jamie
>> Scott wrote: >> >>> I'm running the latest ncl binary for the intel-
>> mac (downloaded >>> ncl_ncarg-5.1.1.MacOS_i386_9.7.0.tar from
>> ESG), >>> and I'm unable to allocate more than 3.5GB before getting
>> a malloc >>> error. I was allocating 200mb variables to see
>> where >>> the limit was. >>> >>>> ncl(46785) malloc: ***
>> mmap(size=200003584) failed (error code=12) >>>> *** error: can't
>> allocate region >>>> *** set a breakpoint in malloc_error_break to
>> debug >>>> fatal:NclMalloc Failed:[errno=12] >>>> fatal:New: could
>> not create new array:[errno=12] >>>> br> >>> >>> So it appears this
>> binary is 32-bit instead of 64. Is this the >>> case or is there
>> something else wrong? If so, is it possible to >>> make a 64-bit
>> binary? >>> >>> I'm using mac OS X 10.58 on and intel-mac with 8GB
>> memory. >>> Darwin Kernel Version 9.8.0: Wed Jul 15 16:55:01 PDT
>> 2009; >>> root:xnu-1228.15.4~1/RELEASE_I386 i386 i386 >>> >>> >>>
>> Thanks, >>> >>> Jamie Scott >>>
>> _______________________________________________ >>> ncl-talk
>> mailing list >>> List instructions, subscriber options,
>> unsubscribe: >>> http://mailman.ucar.edu/mailman/listinfo/ncl-talk
>> >> > >-------------- next part -------------- >An HTML attachment
>> was scrubbed... >URL: http://mailman.ucar.edu/pipermail/ncl-talk/attachments/20091002/13516713/attachment.html
>> > >------------------------------ > >Message: 2 >Date: Fri, 2 Oct
>> 2009 16:30:43 -0600 >From: Wei Huang <huangwei_at_ucar.edu> >Subject:
>> Re: [ncl-talk] intel-mac binary 32 or 64 bit? >To: Jamie Scott <James.D.Scott_at_noaa.gov
>> > >Cc: ncl-talk_at_ucar.edu >Message-ID: <0F305E77-1805-4CD7-9A67-57E7097FA38C_at_ucar.edu
>> > >Content-Type: text/plain; charset="us-ascii" > >Jamie > >It
>> seems that NCL binary for Mac Intel is now 32-bit. >There should be
>> a way to build it in 64-bit. >As half of the NCL team will be
>> traveling the next 3 weeks, >so that will take us longer to look at
>> this. > >Thanks, > >Wei Huang >huangwei_at_ucar.edu >VETS/CISL
>> >National Center for Atmospheric Research >P.O. Box 3000 (1850
>> Table Mesa Dr.) >Boulder, CO 80307-3000 USA >(303) 497-8924 > > > >
>> > >On Oct 2, 2009, at 3:39 PM, Jamie Scott wrote: > >> Doesn't seem
>> to be any limit to memory. Also "top" indicates I have >> 3.3GB
>> available. >> >>> cputime unlimited >>> filesize unlimited
>> >>> datasize 6144 kbytes >>> stacksize 8192 kbytes >>>
>> coredumpsize 0 kbytes >>> memoryuse unlimited >>> descriptors
>> 256 >>> memorylocked unlimited< /font>maxproc 266 >> >> >> -
>> Jamie >> On Oct 2, 2009, at 3:32 PM, Wei Huang wrote: >> >>> Jamie,
>> >>> >>> Your Mac has 8G memory, does not mean your NCL can use all
>> of them. >>> You may have other program running, such as your dock,
>> web browser, >>> finder, etc. >>> For example, I have firefox
>> running, which takes about 150M. >>> >>> Also, there may be per
>> process limit on your mac, and you may >>> use "limit" command to
>> see if there is a limit on each process. >>> >>> Thanks, >>> >>>
>> Wei Huang >>> huangwei_at_ucar.edu >>> VETS/CISL >>> National Center
>> for Atmospheric Research >>> P.O. Box 3000 (1850 Table Mesa Dr.)
>> >>> Boulder, CO 80307-3000 USA >>> (303) 497-8924 >>> >>> >>> >>>
>> >>> >>> On Oct 2, 2009, at 3:10 PM, Jamie Scott wrote: >>> >>>> I'm
>> running the latest ncl binary for the intel-ma c (downl >>>>
>> g-5.1.1.MacOS_i386_9.7.0.tar from ESG), >>>> and I'm unable to
>> allocate more than 3.5GB before getting a malloc >>>> error. I
>> was allocating 200mb variables to see where >>>> the limit was.
>> >>>> >>>>> ncl(46785) malloc: *** mmap(size=200003584) failed
>> (error code=12) >>>>> *** error: can't allocate region >>>>> ***
>> set a breakpoint in malloc_error_break to debug >>>>>
>> fatal:NclMalloc Failed:[errno=12] >>>>> fatal:New: could not create
>> new array:[errno=12] >>>>> br> >>>> >>>> So it appears this binary
>> is 32-bit instead of 64. & nbsp;Is >>>> something else wrong? If
>> so, is it possible to make a 64-bit >>>> binary? >>>> >>>> I'm
>> using mac OS X 10.58 on and intel-mac with 8GB memory. >>>> Darwin
>> Kernel Version 9.8.0: Wed Jul 15 16:55:01 PDT 2009; >>>>
>> root:xnu-1228.15.4~1/RELEASE_I386 i386 i386 >>>> >>>> >>>> Thanks,
>> >>>> >>>> Jamie Scott >>>>
>> _______________________________________________ >>>> ncl-talk
>> mailing list >>>> List instructions, subscriber options,
>> unsubscribe: >>>> http://mailman.ucar.edu/mailman/listinfo/ncl-talk
>> >>> >> > >-------------- next part -------------- >An HTML
>> attachment was scrubbed... >URL: http://mailman.ucar.edu/pipermail/ncl-talk/attachments/20091002/fe21336b/attachment.html
>> > >------------------------------ > >Message: 3 >Date: Sat, 3 Oct
>> 2009 18:38:22 +0530 >From: Oswald Jason Lobo <ojlobo_at_gmail.com>
>> >Subject: "Height" axis label on left in
>> gsn_csm_pres_hgt >To: ncl-talk_at_ucar.edu >Message-ID: > <f1db919c0910030608n3b5b33bax652cefa333f991ff_at_mail.gmail.com
>> > >Content-Type: text/plain; charset="iso-8859-1" > >Hi, > > >
>> I'm trying to put "Height(Km)" on the left Y axis along with the
>> >tickmarks using gsn_csm_pres_hgt. At present the default is the
>> right Y >axis. Is there any way i can put "Height(Km)" on both the
>> axis? Also i'm not >able to change the font of the "Height(Km)"
>> label. > > >Thanks. >-------------- next part -------------- >An
>> HTML attachment was scrubbed... >URL: http://mailman.ucar.edu/pipermail/ncl-talk/attachments/20091003/bbbe64f3/attachment.html
>> > >------------------------------ > >Message: 4 >Date: Sun, 4 Oct
>> 2009 00:12:55 +0800 (CST) >From: eddycarl <eddycarl_at_126.com>
>> >Subject: Strange problem using FORTRAN code in NCL >To:
>> ncl-talk <ncl-talk_at_ucar.edu> >Message-ID: > <8531955.116541254586375684.JavaMail.coremail_at_bj126app67.126.com
>> > >Content-Type: text/plain; charset="gbk" > >Hi there, >I have a
>> problem using a FORTRAN 77 code to read ASCII data into NCL to
>> plot. The data I try to read is TOMS OMI aerosol index and the
>> FORTRAN 77 code is provided at TOMS ftp site. I followed the steps
>> in http://www.ncl.ucar.edu/Document/Tools/WRAPIT.shtml to write
>> special wrapper text, run WRAPIT and load the shared object and
>> call the routine. This seems to be good. But when I run my NCL
>> script, it didn't show ANY result (However, it is supposed to print
>> the information of the the data (for the "print(var)" command)
>> which is read in, and a plot). NO error information is shown,
>> either. When I add a "write" statement in the fortran code (as you
>> can find in my f code), the data can be shown on the commandline
>> when I run NCL. But still, there is NO result from NCL. I think
>> this indicated that the f subrountine can be called by NCL, but
>> there is some problem in transfering data array into NCL. I am
>> wondering if anyone can help to solve this problem? Thanks! >BW: I
>> used NCL Version 5.1.1 on Linux X86_64 platform. I have tested the
>> ex.f and ex.ncl at WRAPIT webpage. And it can be run successfully.
>> > >The fortran code to read the data: (This code is OK to run
>> separately. And it needs a string as filename input) >c this
>> program reads the daily gridded data for real-time OMI >c data
>> for January 31, 2007. variables lat and lon contain the >c
>> latitudes and longitudes of the center of each of the grid
>> >c cells in the array AI. >
>> >C NCLFORTSTART > subroutine read_aersl_omi(fname,lat,lon,ai)
>> > character*80 header
>> > integer*4 ai
>> > real*4 lat(180), lon(360)
>> > character*(*) fname >
>> dimension ai(360,180) >C
>> NCLEND > >c calculate latitudes and
>> longitudes
>> >c >
>> dlat = 1.0 > do 10
>> i=1,180 > lat(i) = -89.5
>> + (i-1)*dlat >10
>> continue > dlon =
>> 1.0 > do 20
>> i=1,360 > lon(i) = -179.5 +
>> (i-1)*dlon >20
>> continue
>> >c >c open the
>> input file
>> >c >cc
>> open
>> (1,file='L3_aersl_omi_20051231.txt',status='old')
>> > open(1,file=fname,status='old')
>> >c
>> >c read in the header lines
>> >c >
>> read(1,'(a80)') header >
>> read(1,'(a80)') header >
>> read(1,'(a80)') header
>> >c >c read
>> in the data into the array ozone
>> >c > do 30
>> i=1,180 > read(1,'(1x,25i3)')
>> (ai(j,i),j=1,360) >30
>> continue
>> >c
>> >c close the input file
>> >c
>> > close(1)
>> >c >c
>> process/print the ai data >c write(*,*)
>> ai > stop >
>> end > >Part of the NCL code I used: >; read multiple files >
>> load "$NCARG_ROOT/lib/ncarg/nclscripts/csm/gsn_code.ncl" > load
>> "$NCARG_ROOT/lib/ncarg/nclscripts/csm/gsn_csm.ncl" > load
>> "$NCARG_ROOT/lib/ncarg/nclscripts/csm/contributed.ncl" > external
>> OMI "./read_aersl_omi.so" ; fortran shared object produced by
>> WRAPIT > begin > fname = "L3_aersl_omi_20051230.txt"
>> > lat = new(180,float) > lon = new(360,float) >
>> var = new((/180,360/),float) > >
>> OMI::read_aersl_omi(fname,lat,lon,var) > print(lat) >
>> print(lon) > print(var) >; ... ... following codes to plot
>> the data > end >-------------- next part -------------- >An HTML
>> attachment was scrubbed... >URL: http://mailman.ucar.edu/pipermail/ncl-talk/attachments/20091004/72fdf49e/attachment.html
>> > >------------------------------ >
>> >_______________________________________________ >ncl-talk mailing
>> list >ncl-talk@ucar.edu >http://mailman.ucar.edu/mailman/listinfo/ncl-talk
>> > > >End of ncl-talk Digest, Vol 71, Issue 6
>> >***************************************
>>
>>
>> 网易邮箱用户购物独享现金返还
>> _______________________________________________
>> ncl-talk mailing list
>> List instructions, subscriber options, unsubscribe:
>> http://mailman.ucar.edu/mailman/listinfo/ncl-talk
>
> _______________________________________________
> ncl-talk mailing list
> List instructions, subscriber options, unsubscribe:
> http://mailman.ucar.edu/mailman/listinfo/ncl-talk

_______________________________________________
ncl-talk mailing list
List instructions, subscriber options, unsubscribe:
http://mailman.ucar.edu/mailman/listinfo/ncl-talk
Received on Mon Oct 05 2009 - 08:14:28 MDT

This archive was generated by hypermail 2.2.0 : Mon Oct 05 2009 - 13:28:34 MDT