Reading a single numerical value from regexp return

I'm searching though text logs and recoridng the number of errors that occur. I want to take the error number after finding it and save it to a variable from the cell array but can't seem to get numerical value out of the cell aray.

The error log looks like the following

Success

Error 303

Success

Completed

Error 301

File

Here is what my code look like at the moment. I get an error when tring to compare the error code to cell array contents.

A = importdata('history_log_28.txt');

numLines = length(A);

Error_301_Count = 0;

Error_302_Count = 0;

Error_303_Count = 0;

for i= 1:numLines

GrabLine = A(i,:);

Error_Catch = regexp(GrabLine,'Error (\w*)','tokens');

if (Error_Catch{1} == '303')

Error_303_Count = Error_303_Count + 1;

end

end

[Stephen23]

"I want to take the error number after finding it and save it to a variable from the cell array but can't seem to get numerical value out of the cell aray. "

There is no numeric value in that cell array, for the very simple reason that REGEXP does not return numeric values from the input text, it returns text. If you want to convert text to numeric, then you can use the usual approaches, e.g. STR2DOUBLE or SSCANF or the like.

Answer 1

[Stephen23]

Get rid of the loop and process the entire file at once. Line-by-line just makes it more complex.

You can get rid of one layer of cell arrays by using a look-behind instead of the token. That simplifies the post-processing.

Don't reinvent the wheel (binning) when it already exists in one efficient function.

txt = fileread('test.txt')

txt = 
    'Success
     Error 303
     Success
     Completed
     Error 301
     File
     Error 303
     hello world'

tmp = regexp(txt,'(?<=Error\s*)\d+','match');
vec = str2double(tmp) % errors found in the file

vec = 1×3

303 301 303

cnt = histcounts(vec,[301:303,Inf]) % count 301, 302, 303.

cnt = 1×3

1 0 2

[Luke]

Thanks the repsonse Stephen! I was implementing the loop so I could also keep track of where the error occured in the timeline via a counter. The intended purpose would be to measure how long the system ran between errors and to assign an error number to a "time" in order to track if the errors become more frequent the longer the runtime. I haven't gotten to that part yet but would something like that sound feasible using you method listed above?

[Stephen23]

"I was implementing the loop so I could also keep track of where the error occured in the timeline via a counter."

txt = fileread('test.txt')

txt = 
    'Success
     Error 303
     Success
     Completed
     Error 301
     File
     Error 303
     hello world'

[tmp,spl] = regexp(txt,'(?<=Error\s*)\d+','match','split');
% the same in my answer... followed by:
dst  = cellfun(@numel,regexprep(spl,'[^\n]','')) % distances

dst = 1×4

1 3 2 1

DST gives the distance from:

• the start of the text to the 1st error
• from the 1st error to the 2nd error
• from the 2nd error to the 3rd error
• etc etc
• from the Nth error to the start of the remaining text.

Note that effectively counts the newlines, so it incudes the error lines as well. You might find CUMSUM useful.

[Luke]

Thank you for the help! I'll take a look at what I have and figure out how to move forward.