Thread Subject:

find complementary values

This may be quite simple but I seem to have a brain block. arrgh.

I have a matrix of with numbers in the range between 1 and 100. The matrix does not contain all values between 1 and 100 and for some numbers there might be multiple instances. For example 0, 10, 11, 11, 15, ..., and 89.

Is there a(n) (elegant) way to find those numbers between 1 and 100 that are not represented in the matrix? i.e., 1-9, 12-14, ...,90-100.

Any help is welcome.

Cheers,
Norman

"Norman " <nteferle@gmail.com> wrote in message <i77trk$2ih$1@fred.mathworks.com>...
> This may be quite simple but I seem to have a brain block. arrgh.
>
> I have a matrix of with numbers in the range between 1 and 100. The matrix does not contain all values between 1 and 100 and for some numbers there might be multiple instances. For example 0, 10, 11, 11, 15, ..., and 89.
>
> Is there a(n) (elegant) way to find those numbers between 1 and 100 that are not represented in the matrix? i.e., 1-9, 12-14, ...,90-100.

TheSet=1:100;
idx=ismember(TheSet,YourValues);
Misising = TheSet(~idx),

"Norman " <nteferle@gmail.com> wrote in message <i77trk$2ih$1@fred.mathworks.com>...
> This may be quite simple but I seem to have a brain block. arrgh.
>
> I have a matrix of with numbers in the range between 1 and 100. The matrix does not contain all values between 1 and 100 and for some numbers there might be multiple instances. For example 0, 10, 11, 11, 15, ..., and 89.
>
> Is there a(n) (elegant) way to find those numbers between 1 and 100 that are not represented in the matrix? i.e., 1-9, 12-14, ...,90-100.
>
> Any help is welcome.
>
> Cheers,
> Norman

Hi Norman,

help setdiff

B = randsample(100,100,true); % requires Statistics Toolbox
% samples 100 values with replacement from 1:100
C = setdiff(1:100,B);

Wayne

"Wayne King" <wmkingty@gmail.com> wrote in message <i77ume$q87$1@fred.mathworks.com>...
> "Norman " <nteferle@gmail.com> wrote in message <i77trk$2ih$1@fred.mathworks.com>...
> > This may be quite simple but I seem to have a brain block. arrgh.
> >
> > I have a matrix of with numbers in the range between 1 and 100. The matrix does not contain all values between 1 and 100 and for some numbers there might be multiple instances. For example 0, 10, 11, 11, 15, ..., and 89.
> >
> > Is there a(n) (elegant) way to find those numbers between 1 and 100 that are not represented in the matrix? i.e., 1-9, 12-14, ...,90-100.
> >
> > Any help is welcome.
> >
> > Cheers,
> > Norman
>
> Hi Norman,
>
> help setdiff
>
> B = randsample(100,100,true); % requires Statistics Toolbox
> % samples 100 values with replacement from 1:100
> C = setdiff(1:100,B);
>
> Wayne

% I don't think setdiff works the way you think it does:


>> A = [0, 10, 11, 11, 15, 89];
>> B = 0:100;
>> c = setdiff(A, B)

c =

   Empty matrix: 1-by-0

"someone" <someone@somewhere.net> wrote in message <i77v8l$546$1@fred.mathworks.com>...

> % I don't think setdiff works the way you think it does:
>
>
> >> A = [0, 10, 11, 11, 15, 89];
> >> B = 0:100;
> >> c = setdiff(A, B)
>
> c =
>
> Empty matrix: 1-by-0
============

It's because you have the input arguments interchanged.

On 20/09/10 10:40 AM, someone wrote:
> "Wayne King" <wmkingty@gmail.com> wrote in message

>> help setdiff
>>
>> B = randsample(100,100,true); % requires Statistics Toolbox
>> % samples 100 values with replacement from 1:100
>> C = setdiff(1:100,B);

> % I don't think setdiff works the way you think it does:

>>> A = [0, 10, 11, 11, 15, 89];
>>> B = 0:100;
>>> c = setdiff(A, B)
>
> c =
>
> Empty matrix: 1-by-0

You have reversed the order of the parameters relative to what Wayne
suggested. setdiff(A,B) finds the elements of A that are not in B. In
your example, there are no elements of A that are not also in B, so the
result is an empty matrix. In Wayne's example, the larger set is in the
A position, so the result would be the values _not_ sampled.

Walter Roberson <roberson@hushmail.com> wrote in message <bsLlo.99878$st2.46575@newsfe09.iad>...
> On 20/09/10 10:40 AM, someone wrote:
> > "Wayne King" <wmkingty@gmail.com> wrote in message
>
> >> help setdiff
> >>
> >> B = randsample(100,100,true); % requires Statistics Toolbox
> >> % samples 100 values with replacement from 1:100
> >> C = setdiff(1:100,B);
>
> > % I don't think setdiff works the way you think it does:
>
> >>> A = [0, 10, 11, 11, 15, 89];
> >>> B = 0:100;
> >>> c = setdiff(A, B)
> >
> > c =
> >
> > Empty matrix: 1-by-0
>
> You have reversed the order of the parameters relative to what Wayne
> suggested. setdiff(A,B) finds the elements of A that are not in B. In
> your example, there are no elements of A that are not also in B, so the
> result is an empty matrix. In Wayne's example, the larger set is in the
> A position, so the result would be the values _not_ sampled.

OOPS

That explains it!

Thanks

"Norman " <nteferle@gmail.com> wrote in message <i77trk$2ih$1@fred.mathworks.com>...
> This may be quite simple but I seem to have a brain block. arrgh.
>
> I have a matrix of with numbers in the range between 1 and 100. The matrix does not contain all values between 1 and 100 and for some numbers there might be multiple instances. For example 0, 10, 11, 11, 15, ..., and 89.
>
> Is there a(n) (elegant) way to find those numbers between 1 and 100 that are not represented in the matrix? i.e., 1-9, 12-14, ...,90-100.
>
> Any help is welcome.
>
> Cheers,
> Norman
- - - - - - - -
  To put the result in something like the missing-interval form you used, do the following. Let A be a row vector of your given numbers (chosen from among numbers 1 to 100).

 x = [0,sort(A),101];
 p = find(diff(x)>1);
 B = [x(p)+1;x(p+1)-1];

  Each of the elements in the top row of B give the initial number of a missing interval and the corresponding number in the bottom row gives the end of that interval. For example if A = [80 44 46 44], you would get

 B = [ 1 45 47 81
      43 45 79 100]

indicating that intervals 1-43, 45-45, 47-79, and 81-100 are missing in A.

Roger Stafford

Many thanks to Matt and everyone else who helped finding a solution to this.
Excellent!
Norman


"Matt J " <mattjacREMOVE@THISieee.spam> wrote in message <i77ui5$hoe$1@fred.mathworks.com>...
> "Norman " <nteferle@gmail.com> wrote in message <i77trk$2ih$1@fred.mathworks.com>...
> > This may be quite simple but I seem to have a brain block. arrgh.
> >
> > I have a matrix of with numbers in the range between 1 and 100. The matrix does not contain all values between 1 and 100 and for some numbers there might be multiple instances. For example 0, 10, 11, 11, 15, ..., and 89.
> >
> > Is there a(n) (elegant) way to find those numbers between 1 and 100 that are not represented in the matrix? i.e., 1-9, 12-14, ...,90-100.
>
> TheSet=1:100;
> idx=ismember(TheSet,YourValues);
> Misising = TheSet(~idx),

"Norman " <nteferle@gmail.com> wrote in message <i77trk$2ih$1@fred.mathworks.com>...
> This may be quite simple but I seem to have a brain block. arrgh.
>
> I have a matrix of with numbers in the range between 1 and 100. The matrix does not contain all values between 1 and 100 and for some numbers there might be multiple instances. For example 0, 10, 11, 11, 15, ..., and 89.
>
> Is there a(n) (elegant) way to find those numbers between 1 and 100 that are not represented in the matrix? i.e., 1-9, 12-14, ...,90-100.
>
> Any help is welcome.
>
> Cheers,
> Norman

For continuous integers the following indexing trick is likely to be faster than setdiff

SomeNumbers = [2 4 6 4 3 9 6 2] % with repetitions
MyRange = 1:10
Result = MyRange % work on a copy
Result(SomeNumbers) = [] % remove from set

hth
Jos