union

Arguments

a, b

can be:

• vectors or matrices of booleans, doubles, strings, duration or datetime. Sparse boolean or numerical matrices are accepted. The types of a and of b can be distinct but compatible for concatenation.
• table and timeseries are also accepted: a and b must have the same variable names.

orient

orientation of the processing:

• 1 or "r": the union is performed over the rows.
• 2 or "c": the union is performed over the columns.

This option is not supported for table and timeseries.

v

depending on type of a and b:

• row vector, or matrix. v's data type is the type of [a(:) ; b(:)]'s result.
• table or timeseries

ka

row vector of integers: Indices in a of v elements/rows/columns coming from a.

kb

row vector of integers: Indices in b of v remaining elements/rows/columns coming from b.

Description

union(a,b), with a and b vectors or matrices, returns a sorted row vector or matrix which retains the unique entries of [a(:) ; b(:)]. If a and b are table or timeseries, union(a, b) returns the sorted unique rows contained in a and b.

union(a,b,"r") or union(a,b,1)returns the matrix formed by the union of the unique rows of a and b sorted in lexicographic ascending order. In this case matrices a and b must have the same number of columns.

union(a,b,"c") or union(a,b,2)returns the matrix formed by the union of the unique columns of a and b sorted in lexicographic ascending order. In this case matrices a and b must have the same number of rows.

The orient option is not managed for table and timeseries.

[v,ka,kb]=union(a,b) also returns index vectors ka and kb such that v is a sorted combination of the entries a(ka) and b(kb).

Examples

A = [6  7  6 ; 5  8  3 ];
B = [1  7  1  0  6 ];

union(A, B)
[u, ka, kb] = union(A, B)

--> union(A, B)
 ans  =
   0.   1.   3.   5.   6.   7.   8.

--> [u, ka, kb] = union(A, B)
 u  =
   0.   1.   3.   5.   6.   7.   8.
 ka  =
   6.   2.   1.   3.   4.
 kb  =
   4.   1.

A = ["a" "b" "a" "c" "c" "b"
     "b" "c" "a" "b" "c" "c" ];
B = ["b" "a" "c" "c" "b" "a"
     "a" "c" "b" "b" "b" "b" ];

[U, ka, kb] = union(A, B, "c")

--> [U, ka, kb] = union(A, B, "c")
 U  =
  "a"  "a"  "a"  "b"  "b"  "b"  "c"  "c"
  "a"  "b"  "c"  "a"  "b"  "c"  "b"  "c"

 ka  =
   3.   1.   2.   4.   5.

 kb  =
   2.   1.   5.

[F, T] = (%f, %t);
A = sparse([T T F T F T ; F F F F T T ; T F F F F T ]); full(A)
B = sparse([F F T T F F ; T T T T T T ; T F T T T F ]); full(B)

[U, ka, kb] = union(A, B, "c");
issparse(U)
full(U), ka, kb

--> A = sparse([T T F T F T ; F F F F T T ; T F F F F T ]); full(A)
 ans  =
  T T F T F T
  F F F F T T
  T F F F F T

--> B = sparse([F F T T F F ; T T T T T T ; T F T T T F ]); full(B)
 ans  =
  F F T T F F
  T T T T T T
  T F T T T F

--> [U, ka, kb] = union(A, B, "c");
--> issparse(U)
 ans  =
  T

--> full(U), ka, kb
 ans  =
  F F F T T T
  F T T F F T
  F F T F T T

 ka  =
   3.   5.   2.   1.   6.
 kb  =
   1.

union() for duration and datetime:

A = hours([1 7 7 4 3 6 6]);
B = hours([7 5 3 0]);
[u, ka, kb] = union(A, B)

A = datetime("today") + A;
B = datetime("today") + B;
[u, ka, kb] = union(A, B)

union() for table:

A = table([1 7 7 4 3 6 6]', [1:7]');
B = table([7 5 3 0]', [2;2;5;5]);
[u, ka, kb] = union(A, B)

A = table([1;3;4;2], ["d";"c";"f";"h"], [%f;%t;%t; %f], "VariableNames", ["double", "string", "boolean"]);
B = table([2;3;4;1], [%t;%t;%t;%t], ["d";"c";"f";"h"], "VariableNames", ["double", "boolean", "string"]);
[u, ka, kb] = union(A, B)

union() for timeseries:

A = timeseries(hours([1:10])', floor(rand(10,1)*10));
B = timeseries(hours(1:2:15)', floor(rand(8,1)*15));
[u, ka, kb] = union(A, B)

See also

• brackets — Concatenation. Recipients of an assignment. Results of a function
• unique — extracts (and sorts) distinct elements, rows or columns of a matrix

History