Count

Counts the number elements in a specified vector that match a specified criteria.

Description

Count(A,B)

COUNT (A,B)

Counts the number of elements in A that match the criteria specified in B.

Example:

Data = [1 2 3 4 3 2 7 6 9 1 1 2 5 9 9];

Count(Data,'==9')

ans =

     3

• MATLAB release: MATLAB 6.5 (R13)

[Mustafa]

I don't want to rate this function, but thanks for the submission. If nothing else it helped me discover the find method.

[Gergely Öllös]

Please consider the FIND command which is already built in..

length(find(Data==9))

[Gergely Öllös]

 

[Mark]

For me histc(x,unique(x)) did what I wanted. This gives me the frequency of each value that occurs in the vector x.

[olcay onder]

Thanks for help.t's very usefull.

[eddy smith]

If last comments have not shown me how to count using sum or find your script would be very usefull.
Your script allowed me to find a good solution. Thanks to you and to the reviewers

[Anil Ramachandran]

dude, just use size(find(index == i),1) to find the number of times i exists in index.

[Jos van der Geest]

Poor help and worse programming style. This file should be removed.

[Dimitri Shvorob]

Duplicates built-in FIND.

[Fabio Wegmann]

Hi I found the lack of a COUNT command quite astonishing. The problem with the solutions offered here is that they 'deliver more than needed' (i.e. position of the element). On a CPU instruction level, this requires an unneeded write to RAM, where an CPU register increment would suffice for the purpose of COUNT. I made three tests with a vector of 64800 elements:

Option 1:
if 1== 1
tic
sum(listeNZ==8)
toc
end

1603 occurences of 8
elapsed time: 0.0051

Option 2:
if 1== 1
tic;
length(find(listeNZ==8))
toc
end

1603 occurences of 8
elapsed time: 0.0038 !!!!!!!! :o (about 30% less time needed)

just for fun:
Option 3:
if 1==1
sums=0;
tic;
for k=1:64800
if listeNZ(k)==8
sums=sums+1;
end
toc
end

1603 occurences of 8
elapsed time: 0.1401 !!!!!!!! :O (about 40 times slower than option 2)

In conclusion, the times needed are too small to worry about, but a COUNT command should really be included in the standard nevertheless... I am really astonished that the length(find()) combo is faster than the sum(), though...

[R M]

Yes, it is true this is not the most efficent function.

I wrote it before I realised there was a find function - (see the date in the help comments) I didn't bother to change it.

[luverly loony]

could also use the 'find' command coupled with 'size' to do the same thing.
>> size(find(Data==9))

[Denis Gilbert]

The function contains too much computing overhead. It also includes an unnecessary EVAL statement. One will usually be better off using the built-in SUM function together with the power of logical indexing.

Example:

>> Data = [1 2 3 4 3 2 7 6 9 1 1 2 5 9 9];
>> sum(Data == 9)

ans =

3

HTH, Denis.