load columns from file with headlines and dates

date, Value(1) , Value(2) , Value(3) , Value(4) ........,Value(200)

date, ?, ?, ?, ?

2010/10/01 00:00:00, 0 , 0 , 0 , 0 ,....

2010/10/01 01:00:00, 12, 5.4, 0 , 0 ,....

2010/10/01 02:00:00, 7.6, 3.99, 1.18, 0 ,...

2010/10/01 03:00:00, 4.06, 0.0008, 4.05, 6.84,...

2010/10/01 04:00:00, 1.44, 0.0020, 4.096, 8.14,...

Hi!

I have a file like the one above (with headers and dates on the first column). With few columns I'm opening it with importfile('myfile.csv');

However I want to load only from, for example, column 100 to column 101. Any suggestions? I was trying with: dlmread('myfile',',','headlines',2) but doesnt work

Thanks

Answer 1

[Chandrasekhar]

it can be read by using csvread command of matlab

[Ana]

It's not working. the file has headlines and the first column of the files are dates...

[Matt Kindig]

I think this should work using textread():

Format = ['%*s %*s ', repmat('%*f', [1 98]), '%f %f', '%*[\n]'];
data = textread('/path/to/your/file.csv', Format, 'delimiter', ',', 'headerLines', 2);

Answer 2

[Cedric]

Assuming that the date/time column has always the same format, the first value starts at char 22 and you can do something like:

 cols   = [100, 101] ;
 buffer = zeros(1e6, length(cols)) ;         % Prealloc for 1e6 entries.

 fid  = fopen('myFile.csv', 'r') ;
 lCnt = 0 ;

 while (~feof(fid))
    line = fgetl(fid) ;
    if line(1) == 'd',  continue ;  end      % Skip lines starting with 'd'.
    lCnt   = lCnt + 1 ;
    values = textscan(line(22:end), '%f,', Inf) ;
    buffer(lCnt,:) = values{1}(cols) ;
 end

 fclose(fid) ;
 buffer = buffer(1:lCnt,:) ;                 % Truncate to number of records.

Note that you might want to realloc buffer when lCnt hits 1e6 (or whichever value you are using for the prealloc). Also, a TEXTREAD guru could probably do this in one shot using this function.