| Re: Little help on arrays [message #38032] |
Mon, 16 February 2004 20:07 |
mchinand
Messages: 66
Registered: September 1996
|
Member |
|
|
In article <4a097d6a.0402161953.238c2a73@posting.google.com>,
M. Katz <MKatz843@onebox.com> wrote:
> mchinand@midway.uchicago.edu (Mike Chinander) wrote in message news:> If
> you don't want to compare element by element, you could check to see if
> the max and min of
>> the difference of the two arrays are the same:
>>
>> IDL> if max(a-b) eq min(a-b) then print, 'bingo!'
>
> This gives the wrong result if there's a constant difference between
> the arrays. For example.
> a = [3,4,5,3]
> b = [2,3,4,2]
>
> then
> max(a-b) = max([1,1,1,1]) = 1
> yet
> min(a-b) = min([1,1,1,1]) = 1
> So they're equal, but the arrays aren't.
>
> M. Katz
Yes, I realized that after seeing someone else's similar solution which
does the needed check to make sure the max is equal to zero.
--Mike
--
Michael Chinander
m-chinander@uchicago.edu
Department of Radiology
University of Chicago
|
|
|
|
| Re: Little help on arrays [message #38033 is a reply to message #38032] |
Mon, 16 February 2004 19:53  |
MKatz843
Messages: 98
Registered: March 2002
|
Member |
|
|
mchinand@midway.uchicago.edu (Mike Chinander) wrote in message news:> If you don't want to compare element by element, you could check to see if the max and min of
> the difference of the two arrays are the same:
>
> IDL> if max(a-b) eq min(a-b) then print, 'bingo!'
This gives the wrong result if there's a constant difference between
the arrays. For example.
a = [3,4,5,3]
b = [2,3,4,2]
then
max(a-b) = max([1,1,1,1]) = 1
yet
min(a-b) = min([1,1,1,1]) = 1
So they're equal, but the arrays aren't.
M. Katz
|
|
|
|
| Re: Little help on arrays [message #38040 is a reply to message #38033] |
Mon, 16 February 2004 12:13 |
Foldy Lajos
Messages: 268
Registered: October 2001
|
Senior Member |
|
|
On Mon, 16 Feb 2004, Craig Markwardt wrote:
>
> Nuno Oliveira <nmoliveira@fc.ul.pt> writes:
>
>
>> How do I compare one array with another? I want to avoid comparing
>> position per position.
>>
>> IDL> if [1,1] eq [1,1] then print, 'bingo!'
>> % Expression must be a scalar or 1 element array in this context: <BYTE
>> Array[2]>.
>> % Execution halted at: $MAIN$
>
> The other posters have good ideas. My own idiom for this comparison
> is:
>
> if total(abs(X-Y)) EQ 0 then print, 'bingo!'
>
> Craig
>
>
Hi,
I think
if min(x-y, max=max) eq max then print, 'bingo!' ; for integer arrays
is about 20% faster for arrays greater than cache memory.
regards,
lajos
|
|
|
|
| Re: Little help on arrays [message #38043 is a reply to message #38040] |
Mon, 16 February 2004 11:23 |
JD Smith
Messages: 850
Registered: December 1999
|
Senior Member |
|
|
On Mon, 16 Feb 2004 09:27:52 -0600, Craig Markwardt wrote:
>
> Nuno Oliveira <nmoliveira@fc.ul.pt> writes:
>
>
>> How do I compare one array with another? I want to avoid comparing
>> position per position.
>>
>> IDL> if [1,1] eq [1,1] then print, 'bingo!'
>> % Expression must be a scalar or 1 element array in this context: <BYTE
>> Array[2]>.
>> % Execution halted at: $MAIN$
>
> The other posters have good ideas. My own idiom for this comparison
> is:
>
> if total(abs(X-Y)) EQ 0 then print, 'bingo!'
If you have a recent enough IDL, ARRAY_EQUAL() is the way to go,
because it stops the comparison as soon as it determines the arrays
are not equal, and the second array need not be an array, but can be a
scalar. I use this all the time for tricks like:
if ~array_equal(array ge 0,1b)
which efficiently determines if any element of array is not ge 0
(i.e. is negative). I also think it's a little more clear what the
intention is than TOTAL. Both the TOTAL and ARRAY_EQUAL method are
much faster than using WHERE, since they don't have to construct list
of indices. Craig's construct is very good when comparing floats, up
to some tolerance:
if total(abs(X-Y) ge 1.e-6) eq 0 then print, 'Close enough!'
but of course this can also be expressed as:
if array_equal(abs(X-Y) lt 1.e-6,1b) then print,'Close enough!'
which would run somewhat faster.
JD
|
|
|
|
| Re: Little help on arrays [message #38049 is a reply to message #38043] |
Mon, 16 February 2004 07:27 |
Craig Markwardt
Messages: 1869
Registered: November 1996
|
Senior Member |
|
|
Nuno Oliveira <nmoliveira@fc.ul.pt> writes:
> How do I compare one array with another? I want to avoid comparing
> position per position.
>
> IDL> if [1,1] eq [1,1] then print, 'bingo!'
> % Expression must be a scalar or 1 element array in this context: <BYTE
> Array[2]>.
> % Execution halted at: $MAIN$
The other posters have good ideas. My own idiom for this comparison
is:
if total(abs(X-Y)) EQ 0 then print, 'bingo!'
Craig
|
|
|
|
| Re: Little help on arrays [message #38051 is a reply to message #38049] |
Mon, 16 February 2004 07:12 |
Pepijn Kenter
Messages: 31
Registered: April 2002
|
Member |
|
|
Nuno Oliveira wrote:
> How do I compare one array with another? I want to avoid comparing
> position per position.
>
> IDL> if [1,1] eq [1,1] then print, 'bingo!'
> % Expression must be a scalar or 1 element array in this context: <BYTE
> Array[2]>.
> % Execution halted at: $MAIN$
>
You can use the array_equal procedure to compare arrays.
By the way, the result of the expresion [1,1] eq [1,1] is also an array
and can therefor not be used in the IF-statement.
IDL> print, [1,1] eq [1,1]
1 1
Pepijn.
|
|
|
|
| Re: Little help on arrays [message #38052 is a reply to message #38051] |
Mon, 16 February 2004 07:12 |
mchinand
Messages: 66
Registered: September 1996
|
Member |
|
|
In article <c0qli8$42a$1@pegasus.fccn.pt>,
Nuno Oliveira <nmoliveira@fc.ul.pt> wrote:
> How do I compare one array with another? I want to avoid comparing
> position per position.
>
> IDL> if [1,1] eq [1,1] then print, 'bingo!'
> % Expression must be a scalar or 1 element array in this context: <BYTE
> Array[2]>.
> % Execution halted at: $MAIN$
>
If you don't want to compare element by element, you could check to see if the max and min of
the difference of the two arrays are the same:
IDL> if max(a-b) eq min(a-b) then print, 'bingo!'
You might want to check to make sure the arrays are the same size first.
--Mike
--
Michael Chinander
m-chinander@uchicago.edu
Department of Radiology
University of Chicago
|
|
|
|
comp.lang.idl-pvwave archive
Members
Search
Help
Login
Home
