How to index the matrix only the first dimension of the rest

Question

Xuze Liu 2022-8-31

0
链接

此问题的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/1791785-how-to-index-the-matrix-only-the-first-dimension-of-the-rest

编辑： Stephen23 2022-9-1

For example, I have an matrix A=zeros(5,4,3)

And B=A(1,:,:)

Is there any way that I don't have to use the latter two ':'?

Something simple like B=A(1,?) that takes whatever from the second dimension onwards?

0 个评论
显示 -2更早的评论隐藏 -2更早的评论

请先登录，再进行评论。

请先登录，再回答此问题。

Answer 1

Bruno Luong 2022-8-31

1
链接

此回答的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/1791785-how-to-index-the-matrix-only-the-first-dimension-of-the-rest#answer_1038880

在 MATLAB Online 中打开

A = rand(2,3,4,5);
C=repmat({':'},1,ndims(A)-1);
% C=repmat({:},1,100); if you don't want to be depend on A
B = A(1,C{:});

0 个评论
显示 -2更早的评论隐藏 -2更早的评论

请先登录，再进行评论。

Answer 2

Stephen23 2022-9-1

1
链接

此回答的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/1791785-how-to-index-the-matrix-only-the-first-dimension-of-the-rest#answer_1039350

在 MATLAB Online 中打开

This will be quite efficient because none of the data in memory is moved around, only the array header is accessed.

It also works for non-scalar X.

B = A(X,:);
S = size(A);
S(1) = size(B,1);
B = reshape(B,S);

1 个评论
显示 -1更早的评论隐藏 -1更早的评论

Stephen23 2022-9-1

编辑：Stephen23 2022-9-1

在 MATLAB Online 中打开

format compact
A = rand(5,4,3)
A = 
A(:,:,1) =
    0.9656    0.6497    0.2581    0.7417
    0.5392    0.6739    0.5440    0.5115
    0.7184    0.7770    0.1037    0.1437
    0.0460    0.8504    0.7329    0.6772
    0.5472    0.3287    0.3364    0.2696
A(:,:,2) =
    0.0271    0.2944    0.1932    0.9287
    0.1086    0.7801    0.5014    0.3357
    0.5984    0.8400    0.6211    0.6933
    0.0673    0.5048    0.7611    0.2261
    0.9548    0.6398    0.8557    0.5052
A(:,:,3) =
    0.8824    0.8974    0.9947    0.4210
    0.0417    0.5307    0.0642    0.1385
    0.7882    0.7552    0.9825    0.7442
    0.5414    0.1979    0.8633    0.4881
    0.7751    0.2920    0.6514    0.4335
X = 3;
B = A(X,:);
S = size(A);
S(1) = size(B,1);
B = reshape(B,S)
B = 
B(:,:,1) =
    0.7184    0.7770    0.1037    0.1437
B(:,:,2) =
    0.5984    0.8400    0.6211    0.6933
B(:,:,3) =
    0.7882    0.7552    0.9825    0.7442

请先登录，再进行评论。

Answer 3

Walter Roberson 2022-8-31

0
链接

此回答的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/1791785-how-to-index-the-matrix-only-the-first-dimension-of-the-rest#answer_1038870

在 MATLAB Online 中打开

B=A(1,:);

However this will treat it like

temp = reshape(A, size(A, 1), []);
B = temp(1,:);

which would give you a 1 x 12 not a 1 x 4 x 3

2 个评论
显示无隐藏无

Xuze Liu 2022-9-1

Thanks. Any way I can avoid this "reshape" behaviour?

Bruno Luong 2022-9-1

在 MATLAB Online 中打开

@Xuze Liu Walter explains the RHS of

B=A(1,:)

is like reshaping A with all the traling dimension grouped. You don't have to.

请先登录，再进行评论。

Answer 4

Cris LaPierre 2022-8-31

0
链接

此回答的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/1791785-how-to-index-the-matrix-only-the-first-dimension-of-the-rest#answer_1038875

在 MATLAB Online 中打开

It depends what you want the ouput dimensions to be. If you just want them in a vector, then you can just use a single colon.

A=rand(5,4,3)
A = 
A(:,:,1) =

    0.8151    0.1142    0.4037    0.5967
    0.0920    0.2222    0.8144    0.6967
    0.1565    0.9195    0.4080    0.7906
    0.1481    0.9235    0.4610    0.4404
    0.6894    0.7504    0.8451    0.0283


A(:,:,2) =

    0.6298    0.0226    0.3653    0.7651
    0.3858    0.3005    0.2023    0.8461
    0.2826    0.2151    0.7336    0.5432
    0.4013    0.0399    0.1543    0.5064
    0.6406    0.9716    0.4377    0.8008


A(:,:,3) =

    0.4181    0.4029    0.7620    0.6040
    0.7262    0.1018    0.3144    0.6743
    0.1473    0.6745    0.5357    0.1704
    0.5709    0.9167    0.0451    0.7542
    0.9045    0.3479    0.7617    0.6889
A(1,:)
ans = 1×12
    0.8151    0.1142    0.4037    0.5967    0.6298    0.0226    0.3653    0.7651    0.4181    0.4029    0.7620    0.6040