Re: Picking up the correct index values

[Przemek Klosowski]

On 5/27/20 9:54 AM, GK19 wrote:
> you cant combine them. It should be 4 arrays. Can you please help me

Octave/Matlab is great at expressing mathematical relationships over 
large datasets.  This implies that you actually pack your data into few 
large arrays and try to find common operations that apply to all of your 
values.

If you insist on separating your data into distinct variables, you 
deprive yourself of the strength and expressiveness of the language, to 
the point that you might as well use Basic or C. Note how Doug packed 
your data into the t1/t 2 arrays, so that he can probe them for equality.

If you organize your data as an array of measurements v(NorP, 
measurement, transistorNumber), so that v(1,3,2) is the third Nwell 
voltage of the second transistor,

v(:,:,1) = [ 0.0 0.9 1.2; -1.8 -1.2 -1.0]

v(:,:,2) = [ 0.9 0 -0.1 ; -1.2 0.2 0.8]

you are not limited to e.g. number of transistors you measured and 
number of measurements per transistor, and you could access the data e,g 
for the second transistor as v(:,:,2), and get all Nwell voltages as

squeeze(v(1,:,:))

(squeeze() eliminates the extra singleton dimension so that you get a 
regular 2D array)

or get the first measurement for all transistors:

squeeze(v(:,1,:))

Then, you can use the ismember() function :

ismember(squeeze(:,2,3), v)

or just loop and compare.

THis is not necessarily obvious if you are used to every value being in 
its own variable, but getting used to this way of representing data is 
very useful when you're dealing with large data sets.