| Re: type reform [message #54733] |
Fri, 06 July 2007 02:57 |
greg.addr
Messages: 160
Registered: May 2007
|
Senior Member |
|
|
On Jul 4, 10:25 pm, bill.d...@gmail.com wrote:
> On Jul 4, 2:07 pm, greg.a...@googlemail.com wrote:
>
>> Is there any way to reform an array so that its type changes? I want
>> to read a file format which contains a mixture of 2 and 4-byte
>> integers. Rather than pre-build complicated structures to read it, it
>> would be nice to read the whole thing as 2-byte values, cut it up as
>> necessary, and then 'reform' the necessary blocks into 4-byte integers
>> with no processing cost.
>
>> Essentially a way to convert an INT = Array[100] into a LONG =
>> Array[50] ?
>
>> Greg
>
> Yes, look at the documentation for LONG(), FIX(), FLOAT(), and the
> other type conversion routines.
> It may not be obvious that when a second argument is supplied, these
> will perform a bit pattern preserving type cast operation similar to
> that in the C language.
>
> Your case for example:
> I = indgen(100)
> L = long(I, 0, 50)
>
> A couple of other examples on a little-endian host (Intel X86
> processor):
> IDL> b = [0B,0B,128B, 63B]
> IDL> print, float(b)
> 0.00000 0.00000 128.000 63.0000
> IDL> print, float(b,0)
> 1.00000
> IDL> print, byte(1.0,0,4)
> 0 0 128 63
>
> Be careful with byte order if you want your code to be portable to all
> platforms supported by IDL!
> -Bill
That worked, but now I met a new problem. If I read the data as bytes
and try to do something like this to cut off a section to convert to
another type:
subset=data[0:n-1,*]
then I run out of memory (when data=~1/2 GB). I suppose IDL is
building a list of indices (4 bytes each?) for every element to copy.
Is there any more efficient way to do this?
Greg
|
|
|
|
|
|
| Re: type reform [message #54755 is a reply to message #54750] |
Wed, 04 July 2007 13:25  |
bill.dman
Messages: 17
Registered: June 2007
|
Junior Member |
|
|
On Jul 4, 2:07 pm, greg.a...@googlemail.com wrote:
> Is there any way to reform an array so that its type changes? I want
> to read a file format which contains a mixture of 2 and 4-byte
> integers. Rather than pre-build complicated structures to read it, it
> would be nice to read the whole thing as 2-byte values, cut it up as
> necessary, and then 'reform' the necessary blocks into 4-byte integers
> with no processing cost.
>
> Essentially a way to convert an INT = Array[100] into a LONG =
> Array[50] ?
>
> Greg
Yes, look at the documentation for LONG(), FIX(), FLOAT(), and the
other type conversion routines.
It may not be obvious that when a second argument is supplied, these
will perform a bit pattern preserving type cast operation similar to
that in the C language.
Your case for example:
I = indgen(100)
L = long(I, 0, 50)
A couple of other examples on a little-endian host (Intel X86
processor):
IDL> b = [0B,0B,128B, 63B]
IDL> print, float(b)
0.00000 0.00000 128.000 63.0000
IDL> print, float(b,0)
1.00000
IDL> print, byte(1.0,0,4)
0 0 128 63
Be careful with byte order if you want your code to be portable to all
platforms supported by IDL!
-Bill
|
|
|
|
comp.lang.idl-pvwave archive
Members
Search
Help
Login
Home
