Re: [opendx-dev] Problem writing loadable modules which use HDF 4
Hi, I've now added a sample data file and sample opendx program to read the data, using the ImportCOARDShdf module which I wrote myself, using the 'import' source code as a basis. I am using OpenDX versions 4.2.0 and 4.3.0, and hdf version 4.1r5. I compiled HDF myself from the tar file on the NCSA site. HDF was compiled with the -DHAVE_NETCDF option although I don't think this relevant in this case. The data and opendx network are here: ftp.soc.soton.ac.uk /pub/jods/opendx/use_importhdf.tar.gz John Michael Wiltberger wrote: Hi John, Thanks for the code. I hope you don't mind if I ask you a few questions because I'm having the same problem with your modules as I did with mine. Namely SDstart returns -1 when I try to open the file. What versions of HDF and OpenDX are you using? Did you compile either set yourself or use precompiled binaries? Would you mind sending putting a datafile on your ftp site so I can see if the problem is with my data? Thanks, Mike John Stark wrote: Hi, I've written a runtime loadable hdf 4 import module designed for single file SD data - I've put the code on our ftp site: ftp.soc.soton.ac.uk It can be found here: mporthdf.tar.gz The code contains makefiles for IRIX, Solaris, and linux - see importhdf.make.[platform]. The code imports array data from a file, taking a start, stride and edge as vectors. It attempts to assign the axes to X,Y,Z,T in a similar method to the ferret package for netcdf data. There are some hacks in there specifically for our ocean model data sets - sorry. John -- John Stark Applications Programmer Southampton Oceanography Centre, UK Tel. +44 (0)23 8059 6571 e-mail jods at soc . soton . ac . uk -- Michael Wiltberger wrote: Hi All, I'm trying to write a runtime loadable module which uses the SD or Multifile Scientific Data Interface to HDF 4.x. The problem is that if I use the SDstart command to open my file it returns an error code. If I use the DFSD or Single-file Scientific Data Interface I can open the file, but I loose the random access features which make HDF so useful. Has anyone else had any success with writing HDF 4.x modules which use the SD interface? Another option would be to use an outboard module (which worked for 4.1.3), but I'm encountering the previously reported problems with outboard modules for version greater than 4.2. Any assistance is greatly appreciated since I'm growing desperate. Thanks, Mike -- ___ |Michael Wiltberger, Ph.D | | |NCAR/HAO | Life's unfair - | |3450 Mitchel Lane| but the root | |Boulder, CO 80301| password helps! | | | | |Ph: (303)497-1532| VI VI VI | |Fx: (303)497-2180| The editor of the | |email: [EMAIL PROTECTED] | beast | |_|___| -- -- ___ |Michael Wiltberger, Ph.D | | |NCAR/HAO | Life's unfair - | |3450 Mitchel Lane| but the root | |Boulder, CO 80301| password helps! | | | | |Ph: (303)497-1532| VI VI VI | |Fx: (303)497-2180| The editor of the | |email: [EMAIL PROTECTED] | beast | |_|___| -- -- John Stark Applications Programmer Southampton Oceanography Centre Tel. +44 (0)23 8059 6571 e-mail [EMAIL PROTECTED] --
Re: [opendx-dev] Problem writing loadable modules which use HDF 4
Hi, I've written a runtime loadable hdf 4 import module designed for single file SD data - I've put the code on our ftp site: ftp.soc.soton.ac.uk It can be found here: /pub/jods/opendx/importhdf.tar.gz The code contains makefiles for IRIX, Solaris, and linux - see importhdf.make.[platform]. The code imports array data from a file, taking a start, stride and edge as vectors. It attempts to assign the axes to X,Y,Z,T in a similar method to the ferret package for netcdf data. There are some hacks in there specifically for our ocean model data sets - sorry. John -- John Stark Applications Programmer Southampton Oceanography Centre, UK Tel. +44 (0)23 8059 6571 e-mail jods at soc . soton . ac . uk -- Michael Wiltberger wrote: Hi All, I'm trying to write a runtime loadable module which uses the SD or Multifile Scientific Data Interface to HDF 4.x. The problem is that if I use the SDstart command to open my file it returns an error code. If I use the DFSD or Single-file Scientific Data Interface I can open the file, but I loose the random access features which make HDF so useful. Has anyone else had any success with writing HDF 4.x modules which use the SD interface? Another option would be to use an outboard module (which worked for 4.1.3), but I'm encountering the previously reported problems with outboard modules for version greater than 4.2. Any assistance is greatly appreciated since I'm growing desperate. Thanks, Mike -- ___ |Michael Wiltberger, Ph.D | | |NCAR/HAO | Life's unfair - | |3450 Mitchel Lane| but the root | |Boulder, CO 80301| password helps! | | | | |Ph: (303)497-1532| VI VI VI | |Fx: (303)497-2180| The editor of the | |email: [EMAIL PROTECTED] | beast | |_|___| --
Re: [opendx-dev] Problem writing loadable modules which use HDF 4
Hi John, Thanks for the code. I hope you don't mind if I ask you a few questions because I'm having the same problem with your modules as I did with mine. Namely SDstart returns -1 when I try to open the file. What versions of HDF and OpenDX are you using? Did you compile either set yourself or use precompiled binaries? Would you mind sending putting a datafile on your ftp site so I can see if the problem is with my data? Thanks, Mike John Stark wrote: Hi, I've written a runtime loadable hdf 4 import module designed for single file SD data - I've put the code on our ftp site: ftp.soc.soton.ac.uk It can be found here: /pub/jods/opendx/importhdf.tar.gz The code contains makefiles for IRIX, Solaris, and linux - see importhdf.make.[platform]. The code imports array data from a file, taking a start, stride and edge as vectors. It attempts to assign the axes to X,Y,Z,T in a similar method to the ferret package for netcdf data. There are some hacks in there specifically for our ocean model data sets - sorry. John -- John Stark Applications Programmer Southampton Oceanography Centre, UK Tel. +44 (0)23 8059 6571 e-mail jods at soc . soton . ac . uk -- Michael Wiltberger wrote: Hi All, I'm trying to write a runtime loadable module which uses the SD or Multifile Scientific Data Interface to HDF 4.x. The problem is that if I use the SDstart command to open my file it returns an error code. If I use the DFSD or Single-file Scientific Data Interface I can open the file, but I loose the random access features which make HDF so useful. Has anyone else had any success with writing HDF 4.x modules which use the SD interface? Another option would be to use an outboard module (which worked for 4.1.3), but I'm encountering the previously reported problems with outboard modules for version greater than 4.2. Any assistance is greatly appreciated since I'm growing desperate. Thanks, Mike -- ___ |Michael Wiltberger, Ph.D | | |NCAR/HAO | Life's unfair - | |3450 Mitchel Lane| but the root | |Boulder, CO 80301| password helps! | | | | |Ph: (303)497-1532| VI VI VI | |Fx: (303)497-2180| The editor of the | |email: [EMAIL PROTECTED] | beast | |_|___| -- -- ___ |Michael Wiltberger, Ph.D | | |NCAR/HAO | Life's unfair - | |3450 Mitchel Lane| but the root | |Boulder, CO 80301| password helps! | | | | |Ph: (303)497-1532| VI VI VI | |Fx: (303)497-2180| The editor of the | |email: [EMAIL PROTECTED] | beast | |_|___|