another import from textfile question

dear ladies and gentlemen

of course this frequently occuring questions about data insertion are very annoying to you, i can understand that, but have to put my issue in here... =(

i have a textfile with maaaaaaaaany columns (2 rows and 13000 columns). its generated from a labview program. unfortunately i am not able to code a formatspec for that amount of columns. so cannot manage to make the data import working.

until now, i refered to https://de.mathworks.com/help/matlab/import_export/ways-to-import-text-files.html and the doc help library

the data in my text.file look like this. actually they are seperated by \t ...

but e.q. a .....

filename = 'S:\...\M0905_PK6.txt';
delimiterIn = '\t';
M = importdata(filename, delimiterIn);

...only creates a 2x1 cell where all values are consequentivly written in one cell per row ..

if i'm trying things like that....

...
    [A,count] = fscanf(fileID, '%f %f');
...

..how can i handle the formatspec of so many columns.

of course i could transform (transpose, what is my original intention) those files manually with notepad++ e.q. per search and replace but i wanted to script it, maybe additional with a file location prompt later

i'm sending the original file and txtfile enclosed

can someone give me a hint =) ?

kindly regards

Basti

Answer 1

[Stephen23]

The main difficulty with your file is that it uses a decimal comma. MATLAB supports only a decimal point. You can search this forum for "decimal comma" to find various solutions, but essentially it boils down to either:

1. altering the file by swapping comma for point before parsing with MATLAB, or
2. importing as character, swap comma for point, convert to numeric.

The first solution you can figure out yourself.

The second solution can be achieved in many ways, here is one which uses efficient sscanf:

str = fileread('M0905_PK6.txt');
str = strrep(str,',','.');
mat = sscanf(str,'%f');
mat = reshape(mat,[],2).';

Giving:

>> mat(:,1:8) % first eight columns
ans =
   0.23343   0.23347   0.23347   0.23353   0.23343   0.23350   0.23340   0.23350
  -0.19354  -0.19354  -0.19350  -0.19360  -0.19350  -0.19354  -0.19354  -0.19363
>> mat(:,end-7:end) % last eight columns
ans =
   0.23038   0.23038   0.23038   0.23045   0.23051   0.23061   0.23058   0.23067
  -0.19324  -0.19334  -0.19324  -0.19328  -0.19321  -0.19318  -0.19315  -0.19318

[Bass Tea]

hey stephen,

you are brilliant. this is all what i need. i have read yesterday such things like "first import as a string, then repair, then transform" but wasn't able to make it working, as well.

i really really thank you very much. the first method u mentioned, clearly make sense, but i wanted to pass that =) so method 2 is great.

appreciate it. have a good one

Answer 2

[Guillaume]

Your file is a simple tab delimited file that is trivially read by csvread or dlmread or readtable:

M = csvread('M0905_PK6.txt')

[Bass Tea]

hi

i really thank you very much for your hint. while i'm doing so, i got such output:

so it's almost done, besides that there is still the problem with my comma decimal seperation sign. how can i tell matlab to only focuss on the \t with a csvread command. matlab must not use comma as a seperater, only \t

regards

[Guillaume]

Oh! I completely missed that it had commas instead of dots for the decimal separators. It's unfortunate that it's still something that matlab doesn't handle properly.