Boost performace for import txt file

1 次查看（过去 30 天）

显示更早的评论

MZ123 2020-4-28

0
链接

此问题的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/521328-boost-performace-for-import-txt-file

评论： MZ123 2020-4-28

file00.txt

在 MATLAB Online 中打开

Good morning everyone,

I'm importing raw data from an instrument which creates a file called "file00.txt" and reading after reading it continues to append new lines to the file. I used to process the file offline, after all the measurements. The file size is normally around 7.5 MB. Attached a reduced version. Every single measurement takes 17 lines organized with a particular schema. To read the file and organize the data I wrote the following code which does its job, but I'd like to push it for better performance. For the attached file matlab scores "Elapsed time is 1.127454 seconds". Can I improve the time elapsed by 1 order of magnitude? Thanks for your help.

fid = fopen('file00.txt');
TotRow = 0; %Initialization of the variable
tline = fgetl(fid); %Set the variable equal to the first line of the file
while ischar(tline) %Counts the number of total lines in the file
  tline = fgetl(fid);
  TotRow = TotRow + 1;
end
nosm = TotRow/17; %Number of single measurements (17 lines per measurements)
D = cell(nosm,219); %Cell initialization
tic
frewind(fid); %Reset file pointer at the beginning
for i = 1 : nosm
    % 17 lines for a single measurements take 1370 different positions of the indicator
    fseek(fid,1370*(i-1),'bof'); %Sets the file position indicator 1370*N bytes from the beginning of the file.   
    CurrentLine = fgetl(fid); %Get 1st line
    D{i,1} = CurrentLine(1:6);
    D{i,2} = CurrentLine(8);
    D{i,3} = CurrentLine(10:23);
    D{i,4} = CurrentLine(28:29);
    D{i,5} = CurrentLine(33:35);
    D{i,6} = CurrentLine(41);
    D{i,7} = CurrentLine(43:48);
    
    CurrentLine = fgetl(fid); %Get 2nd line
    D{i,8} = CurrentLine(2:7);
    D{i,9} = CurrentLine(8:end);
    
    CurrentLine = fgetl(fid); %Get 3rd line, composed of a letter "b" followed by 10 hexadecimal different values
    for j = 1 : 10
        D{i,9+j} = CurrentLine(2+6*(j-1):7+6*(j-1));
    end
    
    for j = 20 : 15 : 200 %Get lines from 4th to 16th, made of a letter followed by 15 hexadecimal different values
        CurrentLine = fgetl(fid);
        for k = 0 : 14
            D{i,j+k} = CurrentLine(2+6*k:7+6*k);
        end
    end
    
    j = 215; %index initialization for last 5 measurements
    CurrentLine = fgetl(fid); %Get last line made of a letter followed by 5 hexadecimal different values
    for k = 0 : 4
        D{i,j+k} = CurrentLine(2+6*k:7+6*k);
    end
 
end
toc