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
andb
must have the same datatype. For text inputs, UTF characters are accepted. Sparse numeric or boolean matrices are accepted : Eithera
orb
or botha
andb
may be sparse. - orient
- flag with possible values : 1 or "r", 2 or "c". Can't be used if
a
or/andb
is an hypermatrix. - M
matrix of the same datatype as
a
andb
.- Without
orient
:M
is a row vector. - With
orient="r"|1
:M
is a matrix stacking the common rows ofa
andb
. - With
orient="c"|2
:M
is a matrix stacking the common columns ofa
andb
.
M
is sparse as soon as eithera
orb
is sparse and none is empty.- Without
- 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""")' 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
History
Versão | Descrição |
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 >> |