| Re: reading past leading and training bytes [message #49201] |
Fri, 07 July 2006 08:01 |
Michael Galloy
Messages: 1114
Registered: April 2006
|
Senior Member |
|
|
Peter Albert wrote:
> While writing this, the idea comes to my head that
> it might have benn smarter to just use
>
> IDL> openw, lun, filename, /get_lun
> IDL> point_lun, lun, 4
>
> instead. Might be worth a try ...
You can even use:
IDL> openw, lun, filename, /get_lun
IDL> skip_lun, lun, 4
to skip 4 bytes. (And this is easier to modify to skip 4 bytes multiple
times instead of just at the beginning of the file.)
Mike
--
www.michaelgalloy.com
|
|
|
|
| Re: reading past leading and training bytes [message #49204 is a reply to message #49201] |
Thu, 06 July 2006 23:47  |
peter.albert@gmx.de
Messages: 108
Registered: July 2005
|
Senior Member |
|
|
Hi Jason,
my previous poster's resonse using the f77_unformatted keyword is most
likely the correct solution to your current problem, but a general
solution for skipping any number of bytes would be:
IDL> openw, lun, filename, /get_lun
IDL> skip_it = bytarr(4) ; replace 4 by the actual number of bytes
you'd like to skip
IDL> readu, lun, skip_it
after that, the file pointer is at the correct position to read the
"real" dataset. Well, that was how I read in those fortran-geenrated
files all the time. While writing this, the idea comes to my head that
it might have benn smarter to just use
IDL> openw, lun, filename, /get_lun
IDL> point_lun, lun, 4
instead. Might be worth a try ...
Cheers,
Peter
glaciologist schrieb:
> I'm using readu to read a 101 X 55 (2D) binary file
>
> The file has leading and trailing 4 byte tags ( 00 00 56 CC in
> hexadecimal).
>
> How to skip past these??!
>
> I use this code, which successfully reads in array, but first column
> gets scrambled!~
>
> readu,1,data
> byteorder,data,/xdrtof ; byteswap needed
>
> data are output by Linux Fedora 4 Portland Group FORTRAN executable
> with -byteswapio flag set on compile
>
> With thanks!
>
> Jason Box
|
|
|
|
| Re: reading past leading and training bytes [message #49217 is a reply to message #49204] |
Thu, 06 July 2006 11:20 |
Paul Van Delst[1]
Messages: 1157
Registered: April 2002
|
Senior Member |
|
|
glaciologist wrote:
> I'm using readu to read a 101 X 55 (2D) binary file
>
> The file has leading and trailing 4 byte tags ( 00 00 56 CC in
> hexadecimal).
>
> How to skip past these??!
>
> I use this code, which successfully reads in array, but first column
> gets scrambled!~
>
> readu,1,data
> byteorder,data,/xdrtof ; byteswap needed
>
> data are output by Linux Fedora 4 Portland Group FORTRAN executable
> with -byteswapio flag set on compile
Without seeing the actual OPEN() statement in your Fortran code (or your IDL code for that
matter), what happens if you use the /f77_unformatted flag in your IDL open? (If you're
not doing it already).
I'm guessing that the 4byte "tags" you're seeing are actually the record markers that
sequential, unformatted Fortran files generally contain to mark the begin and end of
variable length records (with the actual value being the record length).
So try
OPEN, lun, file, /get_lun, /f77_unformatted
You might also need the /swap_endian keyword as well depending on the platform on which
you are reading the byteswapped files. If the linux box that created the files is
little-endian, then the PGI -byteswapio option will produce big-endian files (or vice versa).
paulv
--
Paul van Delst Ride lots.
CIMSS @ NOAA/NCEP/EMC Eddy Merckx
Ph: (301)763-8000 x7748
Fax:(301)763-8545
|
|
|
|
comp.lang.idl-pvwave archive
Members
Search
Help
Login
Home
