Re: READU problems in IDL 4.0.1 for Windows [message #5523] |
Tue, 16 January 1996 00:00  |
Paul R Smith
Messages: 8 Registered: January 1996
|
Junior Member |
|
|
Tom Hamill wrote:
>
> I have been experiencing very odd flaky performance with the READU
> routine in certain circumstances. I use a Fortran program to generate binary
> format files readable by the IDL routine READU. For all files I create, I am
> also able to read them as format binary back into a Fortran program.
> However, in certain situations which do not appear to follow any pattern, the
> IDL routine READU will not be able to fully read some of these binary
> formatted files; it comes back with an error message saying "end of file
> encountered". Does this ring a bell with anyone? Is this a known IDL
> problem, and does anyone have a workaround?
Tom,
I posted a problem a couple of months ago that may be related to what you've found (with only one
response). I am (or was) also using IDL 4.0.1 on Windows 3.1. What I was trying to do was read
structures that contained arrays with largish (in the thousands) numbers of elements from binary
files using READU. The data had been working fine with IDL 4.0 but I got an "end of file" error
when I switched to 4.0.1 The problem for me occurred in the very first read where IDL seemed to
lose several hundred points while reading the first structure and then made up for it by reading
into the next structure. This of couse threw the position out of whack and reading the second
structure failed. It would also mean you would run out of file before your program thought it
had got to the end. I had no idea how to fix this so I switched back to IDL 4.0. I was told a
while later by someone that they had had problems reading large structures in and had broken them
up into smaller reads before combining them into a structure. This would seem to imply that
4.0.1 has trouble reading in chunks of binary data over a certain size but that's a guess. I
haven't experimented further with this though (I just wanted to get on with what I was doing).
Anyone have any more to add about this?
Regards,
Paul
|
|
|