Convert signals of a mf4 file to double variables

Hey guys,

for a project I need to import a MF4 File and for further analysis it needs to be converted to double. It would also be very helpful to predetermine certain variables that are mandatory so that the entire mass of signals doesnt to be processed.

time should also be available as a variable.

as far as of now i got this far:

messdaten=file;

m=mdf(messdaten);

lengthChannelNames = length(m.ChannelNames);

data = cell(1, lengthChannelNames);

for k = 1:lengthChannelNames

data{k} = read(m, k, m.ChannelNames{k});

end

Answer 1

[Cris LaPierre]

Perhaps this example is helpful?

https://www.mathworks.com/help/vnt/ug/reading-data-from-mdf-files.html

Answer 2

[Walter Roberson]

https://www.mathworks.com/help/vnt/ug/read.html#bvcxtko-1

OutputFormat vector

Have you considered just making a single call rather than a loop? It would return a cell of timetable objects. You could cellfun(@(C)C{:,:}, data, 'UniformOutput', 0) to get a cell array of numeric data

[Dennis]

thanks for your answer. is C the cell? I need each signal as double matrix so that i can further process the data. It should have the same functionality as the matlab export feature in canape.

so that the file is being transformed to .mat-file

any idea?

[Walter Roberson]

If you had a cell array of timetable objects that is named YourCellOfTT then you could use

temp = cellfun(@timetable2table, YourCellOfTT, 'UniformOutput', 0);
YourCellOfNumeric = cellfun(@table2array, temp, 'UniformOutput', 0);

to get a cell array of numeric arrays.

Or, more efficiently, you could

YourCellOfNumeric = cellfun(@(DummyParameterNameForCellfunPurposes) DummyParameterNameForCellfunPurposes{:,:}, 'UniformOutput', 0);

which I abbreviated to cellfun(@(C)C{:,:}, data, 'UniformOutput', 0) above

That is, C is a dummy parameter that will be substituted by the contents of each element of the proposed cell array of timetable objects that would result from calling vnt read() to read all of the channels in a single call instead of looping reading one channel at a time.

[Dennis]

Hey thanks!

and what if I just need certain signals? How can I filter for the signals/channel names that are located in different channel groups? and get the values in double format?

[Dennis]

Thank you!

Could you tell me how to search in the cell of tables for certain strings and extract the values?

for instance searching for "kilometer" and extracting the corresponding time and values for the signal

[Walter Roberson]

kilometer_subset = cellfun(@(T) T(strcmp(T.VARIABLE, 'kilometer'),:), YourCellOfTimeTables, 'uniform', 0);

[Dennis]

i dont know why, but it it not work with my data.

attached pictures of how the data cell container looks like.

at the end I need a matrix with the values and the name of the signal that should be extracted through the entire cell array named "data"

[Walter Roberson]

Your data 1 6 does not show any strings or categoricals, so I do not understand where you want to look for 'kilometres'

[Dennis]

So I converted the cell array with timetable to a cell array with tables.

Following that I would like to go thru each of the given tables in the cell and search for the variable names that are needed and save them like above in a vector.

For instance d2s should be searched for or accessed and then saved in a vector.

I dont know how to automate it thru each of the tables that are saved in the Cell array named table Cell. Please help!

[Walter Roberson]

When in doubt about how to do something to multiple items.. write a loop. It might give you code that is not as compact as it could be, but it will not necessarily be any slower. For example each time that cellfun() is used, you have to call through an anonymous function, possibly one with captured variables; if you write the equivalent loop code with a plain call to the work function without going through an anonymous function, the result might be faster than cellfun.

Except in some unusual cases involving real-time systems, it is almost always better to get the functionality working and correct first, and only then stress out about making it more compact or faster. As long as you have a version that is working correctly, (A) you might find it is fast enough anyhow and not worth the effort of improving; and (B) you can use the correctly-working version to compare results against proposed optimizations to be sure that the results are the same. Getting all of the "edge conditions" right with straight-forward code is easier than trying to figure out all of the edge conditions with more compact or more vectorized versions of the code. Your primary duty in the great majority of cases is to get the correct answer first, not to get the fastest or smallest version first.