A loop for sorting tables

I have a long excel file that contains values and times of different variables,which I already sorted by the variable name into different tables (the code below).

Since there are 40 Variables that I want to plot. I would like to know if it is possible to write the 5 lines only once and loop it for each and every variable?

thanks a lot for the help!

tblA = sortrows(Table1,{'Variable','Value'});

Variable_1 = tblA(tblA.Variable == 'Variable 1', :);
Variable_1.time = (seconds(Variable_1.time)/(1000));
Duration_Variable_1=Variable_1.time;
Duration_Variable_1.Format='hh:mm:ss';
Variable_1.time = round(Duration_Variable_1);


Variable_2 = tblA(tblA.Variable == 'Variable 2', :);
Variable_2.time = (seconds(Variable_2.time)/(1000));
Duration_Variable_2=Variable_2.time;
Duration_Variable_2.Format='hh:mm:ss';
Variable_2.time = round(Duration_Variable_2);

Variable_3 = tblA(tblA.Variable == 'Variable 3', :);
Variable_3.time = (seconds(Variable_3.time)/(1000));
Duration_Variable_3=Variable_3.time;
Duration_Variable_3.Format='hh:mm:ss';
Variable_3.time = round(Duration_Variable_3);


Variable_4 = tblA(tblA.Variable == 'Variable 4', :);
Variable_4.time = (seconds(Variable_4.time)/(1000));
Duration_Variable_4=Variable_4.time;
Duration_Variable_4.Format='hh:mm:ss';
Variable_4.time = round(Duration_Variable_4);


Variable_5 = tblA(tblA.Variable == 'Variable 5', :);
Variable_5.time = (seconds(Variable_5.time)/(1000));
Duration_Variable_5=Variable_5.time;
Duration_Variable_5.Format='hh:mm:ss';
Variable_5.time = round(Duration_Variable_5);
.
.
.
.

[Stephen23]

"Since there are 40 Variables that I want to plot..."

Then bad data-design will force you into slow, complex, inefficient, buggy code to access your data:

https://www.mathworks.com/matlabcentral/answers/304528-tutorial-why-variables-should-not-be-named-dynamically-eval

Numbering variables like that is a sign that you are doing something wrong. The simple and efficient approach would be to use indexing (in which case your question is trivially answered: yes, by using a loop and indexing).

"I would like to know if it is possible to write the 5 lines only once and loop it for each and every variable? "

I think Steven Lord can answer that best:

https://www.mathworks.com/matlabcentral/answers/609206-how-to-output-and-name-variables-generated-in-for-loop-into-workspace#answer_510006

[Ramo Rafsel]

Do you happen to have an example how to use one of the alternatives to accessing dynamic variable names? I still am trying with a loop that sorts the table of the variables somewhat automatically but no avail.

thanks a lot in advance for the rescue.

Answer 1

[Stephen23]

tblA = sortrows(Table1,{'Variable','Value'});
for k = 1:40 % or whatever the max 'variable' is.
    vnm = sprintf('Variable %d',k);
    idx = tblA.Variable==vnm;
    tbl = tblA(idx,:);
    tbl.time = seconds(Variable_1.time)/1000;
    tmp = tbl.time;
    tmp.Format='hh:mm:ss';
    tbl.time = round(tmp);
    ... whatever else you want to do, e.g. assign back into the original table:
    tblA(idx,:) = tbl;
end

There are probably better ways to do this, e.g. using retime or dateshift instead of converting to seconds and rounding. But without your data, I cannot test any of this.

[Stephen23]

@Omar Lesfar : in your original question you stated that "...there are 40 Variables that I want to plot", so lets just focus on doing that. It appears that your data was obtained by some kind of event-driven logging, which means the different "Variable" values occur with different/irregular time steps. But the data are already sorted by timestamp, so you do not need to sort the data (but you could easily by using sortrows).

For plotting you can simply loop over the "Variable" values, e.g.:

D = './subdir';
mkdir(D) % only required once
S = load('matlabtable.mat');
T = S.HV801000front10Vml2;
U = unique(T.Variable);
for k = 1:numel(U)
	X = T.Variable==U(k);
	plot(T.time(X),T.Value(X))
	xlabel('time')
	title(char(U(k)),'Interpreter','none')
	F = sprintf('%s.png',char(U(k)));
	saveas(gcf,fullfile(D,F))
end

The first image looks like this (I will not post the other 73 images):

[Stephen23]

"Maybe there is another or a short way that does the same thing without using a code with over 150 lines?"

Yes, use a loop just like I showed you in my answer. Every time you create lots and lots of variables with names like E1_ABD_FoamEventCnt you are forcing yourself into writing lots and lots and lots of code. Basically bad data design (lots of separate variables) forces you into writing bad code (lots of copy-paste). Avoid doing that.

"But I am trying to sort the table I added by the name of the tables"

If you look at my last comment the variable U is sorted, and the loop processes them in that order. You could easily add an explicit list for U if you only want a subset of the "Variable" values:

U = {'E1_ABD_FoamEventCnt','E1_LD_ActualStatus',..}

or use some string matching to only parse the "Variable" values that you want (e.g. strncmp, contains, etc.).

"I did this manually, these are the tables that I wanted to be extracted from the original one..."

You can easily split up that table and store them in a cell array:

tblA = sortrows(Table1,{'Variable','Value'});
out = cell(1,40) % preallocate
for k = 1:40
    vnm = sprintf('Variable %d',k);
    idx = tblA.Variable==vnm;
    tbl = tblA(idx,:);
    ... your code
    out{k} = tbl;
end

Although most likely that is not a particularly good use of the table class, which is designed for processing groups of data within one table.

[Ramo Rafsel]

Thanks you so much for the detailed answer and explanation @Sephen Cobeldick. I see what you mean. I'll re-read everything you wrote here in case I missed an information.

The code works and the line about the U is the thing that bothered me actually the whole time, using and plotting only specific variables that I wanna use.

I am thankful for you answer and your great help!

[Stephen23]

@Omar Lesfar: I hope that it helps! Writing good MATLAB code sometime just requires a slightly different way of looking at a problem... and changing our view of something is often the hardest part!

The volunteers on this forum are always happy to help if you have more questions, or want some examples, etc.