| Re: UNsigned Integer Data [message #8134 is a reply to message #8050] |
Thu, 06 February 1997 00:00  |
Struan Gray
Messages: 178
Registered: December 1995
|
Senior Member |
|
|
David Fanning, davidf@dfanning.com writes:
> array = LONG(array) AND 'FFFF'x
>
> Now you have an array of LONG integers, but they
> have the correct unsigned values. There is no way to get
> around the requirement for LONG integers unless your data
> is always between 0 and 2^31-1 or 2147483647
If you have unsigned, 2-byte integers that go:
0000 0000 0000 0000 : 0
1000 0000 0000 0000 : 32768
1111 1111 1111 1111 : 65535
All you really need to do to get signed integers is
twiddle the top bit:
1000 0000 0000 0000 : 2s-comp -32768
0000 0000 0000 0000 : 2s-comp 0
0111 1111 1111 1111 : 2s-comp 32767
Any routines that calculate a real-world value (ie to label
and axis on a graph) need to have an offset part that accounts
for the fact that 'zero' has moved, but for a lot of datasets
that's a more convenient way of handling things anyway.
Thus you can avoid the memory and time penalty of converting
to longs with something like this:
array = temporary(array) xor fix(-32768)
It definitely works faster on our macs, and since some of
our data files are pushing the memory limits anyway it saves
a lot of disk-swapping for the big ones.
Struan
|
|
|
|
comp.lang.idl-pvwave archive
Members
Search
Help
Login
Home
