Create vector only containing consecutive numbers of other vector

Hi, I'm searching for a possibility to build a vector B which only contains the consecutive numbers of vector A. I tried it like this but it is not working. Can someone help me?

A=[2 5 6 7]; B=0; for m=1:length(A)-1 if diff(A)==1 B=A(m); m=m+1; end end

The anwser should look like this: B=[5 6 7]

Answer 1

[Guillaume]

Using just diff and logical operations (no find):

A = [2 5 6 7 10 12 13 20];
dA = diff(A) == 1;
B = A([0 dA] | [dA 0])

This assumes that A is a row vector.

[Janina Koussev]

In my data A was a column vector... Thanks a lot!

Answer 2

[Stephen23]

For large vectors using logical indexing will probably be faster than using find:

>> A = [2,5,6,7];
>> X = diff(A)==1;
>> Y = [false,X] | [X,false];
>> A(Y)
ans =
   5     6     7

[Stephen23]

And Guillaume edited their answer while I was typing... oh well :)

[Guillaume]

I only edited it to add that it assumed the input was a row vector.

It's Star who, for some reason, deleted his own answer (with find) and the attached comments.

[Stephen23]

Aaah... I do apologize! A case of bad memory and mistaken identity.

[Janina Koussev]

I think the deletion of Star Rider's answer was my mistake as I accepted Guillaume's solution;)

How can I build the vector B with the longest consectutive array only, if I have a vector A with multiple consecutive numbers (e.g.[2 5 6 7 10 12 13 20]). In this example the answer should be B= 5 6 7

[Stephen23]

@Janina Koussev: Ask this in a new question.

[Star Strider]

I delete my Answers if they are neither Accepted nor Voted. Many others do as well.