Using textscan with txt

Hi MATLAB Experts,

I am trying to use textscan function to read txt file, which is actually having more than 100,000 rows and 100 columns.

In order to convert data into a matrix in MATLAB, I am using for loops as follows:

fid = fopen(filename,'r');
tmp = textscan(fid,'%s','Delimiter','\n'); % there are many different data format existing (i.e. char/cell/double)

USdata = [];
for n = 2:length(tmp{1})
     temp_data = regexp(tmp{1}{n},'\t','split');
     USdata = [USdata; temp_data];
 end

However, it takes too long since it has more than 100,000 rows. What would be the most efficient way to do it??

Thanks for any help you could provide in advance.

Jake

Answer 1

[John BG]

Try just reading row by row

fid=fopen('example_text.txt')
A={}
tline = fgetl(fid)
while ischar(tline)
    A=[A; tline]
    tline = fgetl(fid)
end
fclose(fid)

If you find this answer of any help solving your question,

please click on the thumbs-up vote link,

thanks in advance John

[e-mail]

[Jake]

Hi John,

I tried the way you suggested above. It ran much faster. However, I want to a cell matrix (# of column by # of raw). How can I do it?

Answer 2

[Walter Roberson]

ncol = 100;
fmt = strjoin(repmat({'%[^\t]'},1,ncol),'\t');
fid = fopen(filename,'r');
tmp = textscan(fid, fmt, 'CollectOutput', 1));
fclose(fid);
data = tmp{1};

If you know the column formats, it would be better to construct a more specific fmt. You might want to adapt some of the methods I use in http://www.mathworks.com/matlabcentral/answers/280772-how-can-i-load-target-data-on-student-alcohol-consumption#answer_219273