fast and beautiful way to convert vector ranges to indexes

Question

Hi there,   my question could easily be solved with a simple loop but I'm curious if there is a nice genuine matlab way of doing this: I have a vector "v1" containing 10 entries. First 5 belong together, next 2 belong together and the last 3 again (somehow). So I got a vector "v2" with     v2 = [5; 2; 3]; I want now something like v3 = [1 1 1 1 1 2 2 3 3 3]; So I could access my v1 vector with: v1(v3==1); (Background for the question are different colors for each group with the plot-command.)   Thanks already in advance - I'm sure Matlab holds a nice and short way of doing this :-)

Kshitij Singh · Accepted Answer

c=cumsum(v2);
v3=zeros(1,c(end));
 v3([1,c(1:end-1)])=1;

v3=cumsum(v3);
 v3(end)=[],
Summary:


The cumsum(v2) creates a cumulative sum vector c.


A zero vector v3 is created with a length of the last value in c.


Positions in v3 at 1 and c(1:end-1) are set to 1.


Another cumulative sum is calculated over this v3.


The last element of v3 is removed.


The result is a vector v3 that mirrors the cumulative structure of v2 but adjusted for the specific indexing manipulation.
This process transforms a cumulative indexing approach into a sequential one, useful in various signal processing or data transformation scenarios.

fast and beautiful way to convert vector ranges to indexes

Summary:

Get a Free Consultation or a Sample Assignment Review!