Convert table format from %s to %f

I put together a table with variables that I imported from an excel file and now I am unable to plot them.

I think I need to convert the format of the values from %s to %f, but I am unsure how.

Here is some code that I have:

filename = 'filename'; %Name of file to be imported
[num,txt,raw] = xlsread(filename);

% Universal data grab
[rTime,cTime] = find(strcmp(raw,'Time')); %Top of table (Headers)
NanInd = find(isnan(num(:,1)));
rawNan = [NanInd(1)+rTime-1,1]; % Bottom of table

% Create table
headers = raw(rTime,:);
nHeaders = char(headers);
Headers = strrep(headers,')','_');
Headers = strrep(Headers,'(','_');
Headers = strrep(Headers,'-','_');
Headers = strrep(Headers,'-','_');
Headers = strrep(Headers,' ','_');
Headers = strrep(Headers,'_H','x_T');

T = array2table(raw(rTime+1:rawNan(1),:),'VariableNames',Headers);
T.Time = datetime(datevec(cell2mat(T.Time)),'Format','HH:mm:ss.SSS');

plot(T.Time,T.VarY,'k')

This is the error I get:

"Error using datetime/plot (line 78)

Invalid second data argument"

Can I only plot if my data is format %f and how can I convert to %f?

[Guillaume]

Why aren't you using readtable if your excel sheet has a header?

Also, you could use matlab.lang.makeValidName to create valid variable name (readtable would do that for you) instead of all these strrep.

An example excel file would be useful for us to understand what the problem is.

[Ryan McBurney]

The excel file is an output from data collection and depending on what the data collected is there is several lines/cells above the actual data I want in the excel file so I search for the word Time to find the starting point. And then find the emtpy (NaN) cell at teh bottom of the data I want.

I used readtable originally, but I would have to change the range for different kinds of data sets (excel files). I thought what I did would make the code more applicable to different excel files. But now I have my current issue.

I can't show the data/excel file unfortunately.

Thanks for the heads up with matlab.lang.makeValidName. That's a new one for me.

[dpb]

Don't use the raw data to make the numeric table entries--use the num array instead. xlsread will have already done the numeric conversion. Use the raw data only for the purpose of finding the data locations as have done. (I've done this precise exercise a lot for a bunch of Excel files in use at the college where I'm doing pro bono work...sometimes it's easier to read and find the data range then use that with readtable

You may also find that if you create a spreadsheet import object it may have enough additional flexibility that readtable alone doesn't have...altho you can't find a variable number of headerlines that way always reliably depending on the file structure.

If you can't attach an example file or two, we're pretty much only able to give generic advice rather than explicit soutions, unfortunately.

[Ryan McBurney]

You nailed it, dqp. I needed to grab the Time variable from the raw data output and num data for the rest of my variables.

tempT = array2table(raw(rTime+1:rawNan(1),:),'VariableNames',Headers);
tempT.Time = datetime(datevec(cell2mat(tempT.Time)),'Format','HH:mm:ss.SSS');

T = array2table(num(1:NanInd(1)-1,:),'VariableNames',Headers);
T.Time = tempT.Time;

Thanks again Guillaume matlab.lang.makeValidName. Definitely makes my code more applicable to other excel files.