Split array according to one column values

I have an array like this:
A = {10 1 1; 10 1 2; 10 1 1; 10 2 1; 10 1 3; 10 2 4; 10 2 5}
I want to split into arrays based on the second column value. The output should look like this:
B = {10 1 1; 10 1 2; 10 1 1; 10 1 3}
C = {10 2 1; 10 2 4; 10 2 5}
How can I do this ?

2 Kommentare

Jack Lavender
Jack Lavender am 22 Aug. 2022
What would you recommend if you have a large array and an unknown number of different values in the indexing column?

Melden Sie sich an, um zu kommentieren.

Antworten (3)

Stephen23
Stephen23 am 25 Jul. 2018
Bearbeitet: Stephen23 am 25 Jul. 2018

2 Stimmen

Your life would be much easier if you stored numeric data in numeric arrays:
>> A = [10 1 1; 10 1 2; 10 1 1; 10 2 1; 10 1 3; 10 2 4; 10 2 5]
A =
10 1 1
10 1 2
10 1 1
10 2 1
10 1 3
10 2 4
10 2 5
>> B = A(A(:,2)==1,:)
B =
10 1 1
10 1 2
10 1 1
10 1 3
>> C = A(A(:,2)==2,:)
C =
10 2 1
10 2 4
10 2 5

4 Kommentare

Rui Carapinha
Rui Carapinha am 25 Jul. 2018
Thanks for the fast response :)
What I represented has a example, I can have several values in the 2nd column (sorry if I didn't explained this). That way I would have to know in the first hand the values I have.
@Rui Carapinha: there are several approaches. Here is one simple solution using unique and arrayfun:
>> A = [10,1,1;10,1,2;10,1,1;10,2,1;10,1,3;10,2,4;10,2,5]
A =
10 1 1
10 1 2
10 1 1
10 2 1
10 1 3
10 2 4
10 2 5
>> U = unique(A(:,2));
>> C = arrayfun(@(n)A(A(:,2)==n,:),U,'uni',0);
>> C{:}
ans =
10 1 1
10 1 2
10 1 1
10 1 3
ans =
10 2 1
10 2 4
10 2 5
Stephen23
Stephen23 am 22 Aug. 2022
Bearbeitet: Stephen23 am 22 Aug. 2022
Another approach using FINDGROUPS and SPLITAPPLY:
A = [10,1,1;10,1,2;10,1,1;10,2,1;10,1,3;10,2,4;10,2,5]
A = 7×3
10 1 1 10 1 2 10 1 1 10 2 1 10 1 3 10 2 4 10 2 5
G = findgroups(A(:,2));
C = splitapply(@(m){m},A,G)
C = 2×1 cell array
{4×3 double} {3×3 double}
C{:}
ans = 4×3
10 1 1 10 1 2 10 1 1 10 1 3
ans = 3×3
10 2 1 10 2 4 10 2 5
Steven Lord
Steven Lord am 23 Aug. 2022
You could potentially also use groupsummary and use it to perform whatever operations you'd perform on the matrices stored in the cells of C.

Melden Sie sich an, um zu kommentieren.

LuKr
LuKr am 25 Jul. 2018

1 Stimme

B=A(cellfun(@(x) isequal(1,x),{A{:,2}}),:)
C=A(cellfun(@(x) isequal(2,x),{A{:,2}}),:)
Maybe like this?
Hermes Suen
Hermes Suen am 25 Jul. 2018

0 Stimmen

You should be able to do this through logical indexing, and it may be easier if it was not a cell array but a normal numerical array. Define the matrix as such, without the curly braces:
A = [10 1 1; 10 1 2; 10 1 1; 10 2 1; 10 1 3; 10 2 4; 10 2 5];
Then store the second column into a separate variable:
cols = A(:,2);
This takes the second column of the A matrix and stores it into a column vector called "cols". From there you can use logical indexing as follows:
ones = cols ==1;
twos = cols ==2;
The output of the above two commands should be this:
ones =
7×1 logical array
1
1
1
0
1
0
0
"Ones" will have a "1" whenever the column vector has a "1" and has a "0" whenever the column vector has anything other than a 1. The same logic then applies to the variable "twos"
From here you can then call the following two commands to get what you are looking for:
B = A(ones,:)
C = A(twos,:)
This is logical indexing, a neat feature of MATLAB, let me know if you need further clarification or if you have any questions!

Kategorien

Gefragt:

am 25 Jul. 2018

Kommentiert:

am 23 Aug. 2022

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!

Translated by