Scilab Website | Contribute with GitLab | Mailing list archives | ATOMS toolboxes
Scilab Online Help
2024.1.0 - English


intersect

elements or rows or columns met in both input arrays, without duplicates

Syntax

M = intersect(a, b)
M = intersect(a, b, orient)
[M, ka] = intersect(..)
[M, ka, kb] = intersect(..)

Arguments

a, b
vectors, matrices or hypermatrices of booleans, encoded integers, real or complex numbers, or text. a and b must have the same datatype. For text inputs, UTF characters are accepted. Sparse numeric or boolean matrices are accepted : Either a or b or both a and b may be sparse.

orient
flag with possible values : 1 or "r", 2 or "c". Can't be used if a or/and b is an hypermatrix.

M

matrix of the same datatype as a and b.

  • Without orient: M is a row vector.
  • With orient="r"|1: M is a matrix stacking the common rows of a and b.
  • With orient="c"|2: M is a matrix stacking the common columns of a and b.

M is sparse as soon as either a or b is sparse and none is empty.

ka
Dense row vector of indices in a.

kb
Dense row vector of indices in b.

Description

intersect(a,b) returns a row vector of unique values present in both a and b arrays. Values are sorted in increasing order

  • for complex numbers : par increasing magnitudes, then by increasing phases
  • for text : in alphabetical order.

Two NaN elements are always considered as different. So NaN or rows or columns having NaN will never be in the result M.

[M, ka, kb] = intersect(a,b) additionally returns the vectors ka and kb of indices in a and b of selected components firstly met, such that M=a(ka) and M=b(kb).

Common rows or columns

When the orient argument is provided, the comparison is performed between the rows of a and b -- each one being considered as a whole --, or between their columns.

intersect(a,b,"r") or intersect(a,b,1) will return the matrix of stacked unduplicated rows met in both a and b, sorted in lexicographic ascending order. If a and b don't have the same number of columns, [] is returned without comparing the values.

[M,ka,kb] = intersect(a,b,"r") additionally returns the vectors ka and kb of the minimal indices of common rows, respectively in a and b, such that M=a(ka,:) and M=b(kb,:).

intersect(a,b,"c") or intersect(a,b,2) does the same for columns.

Examples

A = grand(3, 3, "uin", 0, 9)
B = grand(2, 4, "uin", 0, 9)
intersect(A, B)
[N, ka, kb] = intersect(A,B);
ka, kb
--> A = grand(3, 3, "uin", 0, 9)
 A  =
   0.   6.   4.
   6.   6.   6.
   2.   7.   9.

--> B = grand(2, 4, "uin", 0, 9)
 B  =
   1.   8.   0.   2.
   6.   2.   2.   1.

--> intersect(A, B)
 ans  =
   0.   2.   6.

--> [N, ka, kb] = intersect(A,B);
--> ka, kb
 ka  =
   1.   3.   2.
 kb  =
   5.   4.   2.

In the above example, note that 6 is met four times in A, at indices [2 4 5 8]. Only the minimal index 2 is returned in ka. Same situation for 2 in B.

NaN values can never be in the result:

%nan == %nan
intersect([1 -2 %nan 3 6], [%nan 1:3])
--> %nan == %nan
 ans  =
  F

--> intersect([1 -2 %nan 3 6], [%nan 1:3])
 ans  =
   1.   3.

intersect() can also process some characters or some texts. Since Scilab is great with UTF characters, here is an example with some Arabic contents, getting characters present in both sentences:

A = strsplit("هو برنامج علمي كبير ""Scilab""")'
B = strsplit("فهو حر ومفتوح")'
intersect(A,B)
--> A = strsplit("هو برنامج علمي كبير ""Scilab""")'
 A  =
!ه  و     ب  ر  ن  ا  م  ج     ع  ل  م  ي     ك  ب  ي  ر     "  S  c  i  l  a  b  "  !

--> B = strsplit("فهو حر ومفتوح")'
 B  =
!ف  ه  و     ح  ر     و  م  ف  ت  و  ح  !

--> intersect(A,B)
 ans  =
!   ر  م  ه  و  !

Column-wise or Row-wise processing of two matrices: Here we process 2 matrices of signed 1-byte integers, and get the common columns:

A = int8(grand(3,5,"uin",0,1))
B = int8(grand(3,9,"uin",0,1))
[M,ka,kb] = intersect(A, B, "c");
M, ka, kb
--> A = int8(grand(3,5,"uin",0,1))
 A  =
  0  0  1  1  1
  0  0  1  1  0
  0  0  0  0  1

--> B = int8(grand(3,9,"uin",0,1))
 B  =
  1  0  1  1  1  0  1  1  1
  1  0  0  1  1  1  0  0  0
  1  0  1  0  1  1  1  0  0

--> [M,ka,kb] = intersect(A, B, "c");
--> M, ka, kb
 M  =
  0  1  1
  0  0  1
  0  1  0

 ka  =
   1.   5.   3.

 kb  =
   2.   3.   4.

intersect() for booleans is mainly useful with the "r" or "c" option. Here is an example with a sparse boolean matrix:

[F, T] = (%f, %t);
A = [F F T F T F ; T F F T T T ; T T F T F F]
B = [F T F T F F ; T F F F T F ; F T F F T F]
[M,ka,kb] = intersect(A, sparse(B), "c");
issparse(M), full(M), ka, kb
--> A = [F F T F T F ; T F F T T T ; T T F T F F]
 A  =
  F F T F T F
  T F F T T T
  T T F T F F

--> B = [F T F T F F ; T F F F T F ; F T F F T F]
 B  =
  F T F T F F
  T F F F T F
  F T F F T F

--> [M,ka,kb] = intersect(A, sparse(B), "c");
--> issparse(M), full(M), ka, kb
 ans  =
  T

 ans  =
  F F T
  T T F
  F T F

 ka  =
   6.   1.   3.

 kb  =
   1.   5.   4.

See also

  • members — count (and locate) in an array each element or row or column of another array
  • unique — extracts (and sorts) distinct elements, rows or columns of a matrix
  • gsort — sorts boolean, numerical and string arrays
  • union — Set of all elements, rows, or columns of two arrays, without duplicates

History

VersionDescription
6.1.0 Complex numbers are now accepted.
6.1.1 Sparse matrices are now accepted (numbers or booleans).
Report an issue
<< Set operations Set operations setdiff >>

Copyright (c) 2022-2024 (Dassault Systèmes)
Copyright (c) 2017-2022 (ESI Group)
Copyright (c) 2011-2017 (Scilab Enterprises)
Copyright (c) 1989-2012 (INRIA)
Copyright (c) 1989-2007 (ENPC)
with contributors
Last updated:
Mon Jun 17 17:49:15 CEST 2024