Partitioning a vector of doubles into cells using a second vector to index

Hi all,

I would like to partition a vector of numbers into separate sub-cells, using a second vector as a list of indices. See the example below:

vectorContents = 1:5;
indexingVector = [1 1 1 2 2];

I'd like the output to be a cell array that looks like

storeCell = {[1 1 1],[2 2]}

Is there a way to achieve this using deal and other parallelized functions to avoid a for loop? My current implementation looks like this:

vectorContents = 1:5;
indexingVector = [1 1 1 2 2];
uniqueInds = unique(indexingVector);
storeCell = cell(numel(uniqueInds),1);
for ii = 1:numel(uniqueInds)
    
    currInd = uniqueInds(ii);
    currLogic = indexingVector == currInd;
    storeCell{ii} = vectorContents(currLogic);
    
end

Thanks in advance!

Answer 1

[Stephen23]

V = 1:5;
X = [1,1,1,2,2];
C = accumarray(X(:),V(:),[],@(v){v.'})

C = 2x1 cell array
    {[1 2 3]}
    {[  4 5]}

or

C = groupsummary(V(:),X(:),@(v){v.'})

C = 2x1 cell array
    {[1 2 3]}
    {[  4 5]}

or

C = arrayfun(@(n)V(n==X),1:max(X),'Uni',0) % probably the slowest

C = 1x2 cell array
    {[1 2 3]}    {[4 5]}

[Remi Boros]

Thank you for your rigorous answer, despite my typo

Answer 2

[Pavl M.]

clc
clear all
close all

vectorContents = 1:7

vectorContents = 1×7

1 2 3 4 5 6 7

<mw-icon class=""></mw-icon>

<mw-icon class=""></mw-icon>

indexingVector = [1 1 1 2 2 3 3]

indexingVector = 1×7

1 1 1 2 2 3 3

<mw-icon class=""></mw-icon>

<mw-icon class=""></mw-icon>

uniqueInds = unique(indexingVector)

uniqueInds = 1×3

1 2 3

<mw-icon class=""></mw-icon>

<mw-icon class=""></mw-icon>

storeCell = cell(numel(uniqueInds),1)

storeCell = 3x1 cell array
    {0x0 double}
    {0x0 double}
    {0x0 double}

%Solution with no loops:

tic
%Form vector of distinct values repetitions (acc to your custom logic):
vec = [3 2 2]; % 1D vector
storeCell = mat2cell(vectorContents,1,vec)

storeCell = 1x3 cell array
    {[1 2 3]}    {[4 5]}    {[6 7]}

toc

Elapsed time is 0.085405 seconds.

%Ver of Stephen23 measurements

tic
C = accumarray(indexingVector(:),vectorContents(:),[],@(vectorContents){vectorContents.'})

C = 3x1 cell array
    {[1 2 3]}
    {[  4 5]}
    {[  6 7]}

toc

Elapsed time is 0.053025 seconds.

tic
C = arrayfun(@(n)vectorContents(n==indexingVector),1:max(indexingVector),'Uni',0) % probably the slowest

C = 1x3 cell array
    {[1 2 3]}    {[4 5]}    {[6 7]}

toc

Elapsed time is 0.162446 seconds.

tic
C = groupsummary(vectorContents(:),indexingVector(:),@(indexingVector){indexingVector.'})

C = 3x1 cell array
    {[1 2 3]}
    {[  4 5]}
    {[  6 7]}

toc

Elapsed time is 0.156229 seconds.

%Solution with 1 loops:
tic
for ii = 1:numel(uniqueInds)

    currInd = uniqueInds(ii);
    currLogic = indexingVector == currInd;
    storeCell{ii} = vectorContents(currLogic)

end

storeCell = 1x3 cell array
    {[1 2 3]}    {[4 5]}    {[6 7]}

storeCell = 1x3 cell array
    {[1 2 3]}    {[4 5]}    {[6 7]}

storeCell = 1x3 cell array
    {[1 2 3]}    {[4 5]}    {[6 7]}

toc

Elapsed time is 0.059105 seconds.

%Constructed by P.Mazniker
    %Constructed from needing help code by
    %https://join.skype.com/invite/oXnJhbgys7oW
    %https://independent.academia.edu/PMazniker
    %[phone]
    %https://diag.net/u/u6r3ondjie0w0l8138bafm095b
    %https://github.com/goodengineer
    %https://orcid.org/0000-0001-8184-8166
    %https://willwork781147312.wordpress.com/portfolio/cp/
    %https://www.youtube.com/channel/UCC__7jMOAHak0MVkUFtmO-w
    %https://nanohub.org/members/130066
    %https://pangian.com/user/hiretoserve/
    %https://substack.com/profile/191772642-paul-m

[Remi Boros]

Thank you for your rigorous answer, despite my typo