Matlab - importing ascii data - matlab

Question

I ve got an ascii file and im trying to import it to matlab in order to make some plots. Is there any way of importing those data, even tho they contain , (comma) rather than . (dot)?
00:00:00,000;-2,14;
00:00:00,001;-1,80;
Well the first column which I want to create is referred to the time and its corresponding to 00:00:00,001; 00:00:00,002; etc.
The second column should be the amplitude of the sample i.e. -2,14; -1,80 etc.

Answer 1

Yup. First use importdata so that you can read each row of your text file as a cell in a cell array. After, to allow for the processing of your times to be performed in MATLAB, you'll need to replace each , character with a .. This will allow you to use MATLAB's commands for date and time processing. Specifically, use regular expressions to help you do this. Regular expressions help you find patterns in strings. We can use these patterns to help extract out the data you need. Use regexprep to replace all , characters with a ..
For the purposes of this answer, the example data that I'm going to be using is:
00:00:00,000;-2,14;
00:00:00,001;-1,80;
00:00:00,002;-0,80;
00:00:00,003;2,40;
00:00:00,004;3,78;
Therefore, assuming that your data is stored in a text file called data.txt, do:
%// Load in each row as a cell array
A = importdata('data.txt');
%// Each row has , replaced with .
Arep = regexprep(A, ',', '\.');
Now, what we can do is split up all of the quantities by using ; as the delimiter. We can use regexp to help us split up the quantities. We can further decompose the data by:
Arep_decomp = regexp(Arep, '[^;]+', 'match');
The first parameter is the cell array that contains each of our rows in the text file (with the commas converted to periods). The second parameter is a pattern that specifies what exactly you're trying to look for in each string in the cell array. [^;]+ means that you want to find all strings that consist of a bunch of characters excluding until we hit a semi-colon. Once we hit the semi-colon, we stop. 'match' means that you want to retrieve the actual strings which will be stored as cell arrays.
The result after the above line's execution gives:
Arep_decomp{1}{1} =
00:00:00.000
Arep_decomp{1}{2} =
-2.14
Arep_decomp{2}{1} =
00:00:00.001
Arep_decomp{2}{2} =
-1.80
Arep_decomp{3}{1} =
00:00:00.002
Arep_decomp{3}{2} =
-0.80
Arep_decomp{4}{1} =
00:00:00.003
Arep_decomp{4}{2} =
2.40
Arep_decomp{5}{1} =
00:00:00.004
Arep_decomp{5}{2} =
3.78
You can see that the output cell array, Arep_decomp is a 5 element cell array, where each cell is a nested 2 element cell array, where the first element is the time, and the second element is the magnitude. Note that these are all strings.
What you can do now is create two numeric arrays that will convert these quantities into numeric representations. Specifically, the time format that you have looks like the form:
HH:MM:SS.FFF
H is for hours, M is for minutes, S is for seconds and F is for microseconds. Use datenum to allow you to convert these time representations into actual date numbers. You would do this so that you can plot these on a graph, but then you perhaps want to display these times on the plot as well. That can easily be done by manipulating some plot functions. Nevertheless, use cellfun so that we can extract out the time strings as a separate array so we can use this for plotting later, and also use this to convert the time strings into date numbers via datenum, and convert the magnitude numbers into actual numbers.
Therefore:
datestr = cellfun(#(x) x{1}, Arep_decomp, 'uni', 0);
datenums = cellfun(#(x) datenum(x, 'HH:MM:SS.FFF'), datestr);
mags = cellfun(#(x) str2double(x{2}), Arep_decomp);
The first line of code extracts out each of the time strings as a single cell array - the uni=0 flag is important to do this. Next, we convert each time string into a date number, and we convert the magnitude strings into physical numbers by str2double.
Now, all you have to do is plot the data. That can be done by:
plot(datenums, mags);
set(gca, 'XTick', datenums);
set(gca, 'XTickLabel', datestr);
The above code plots the data where the date numbers are on the horizontal axis, the magnitude numbers are on the vertical axis, but we will probably want to rename the horizontal axis to be those time strings that you wanted. Therefore, we use to calls to set to ensure that the only ticks that are visible are from the date numbers themselves, and we relabel the date numbers so that they are the string representations of the times themselves.
Once we run the above code, we get:
Because the time step in between times is so small, it may clutter the horizontal axis as the labels are long, yet the interval is short. Therefore, you may consider only displaying times at a certain interval and you can do that by doing something like:
step_size = 5;
plot(datenums, mags);
set(gca, 'XTick', datenums(1:step_size:end));
set(gca, 'XTickLabel', datestr(1:step_size:end));
step_size controls how many ticks and labels appear in succession. Obviously, you need to make sure that step_size is smaller than the total number of points in your data.
For your copying and pasting pleasure, this is what the full code I wrote looks like:
%// Load in each row as a cell array
A = importdata('data.txt');
%// Each row has , replaced with .
Arep = regexprep(A, ',', '\.');
Arep_decomp = regexp(Arep, '[^;]+', 'match');
datestr = cellfun(#(x) x{1}, Arep_decomp, 'uni', 0);
datenums = cellfun(#(x) datenum(x, 'HH:MM:SS.FFF'), datestr);
mags = cellfun(#(x) str2double(x{2}), Arep_decomp);
step_size = 1;
%step_size = 5;
plot(datenums, mags);
set(gca, 'XTick', datenums(1:step_size:end));
set(gca, 'XTickLabel', datestr(1:step_size:end));