import data from a .txt file with multiple delimiters and headerlines

Hello dear Matlab Gurus,

I would like to know a way to import data from a .txt file. The file has 5 lines which I would like to ignore and the columns are separated by commas and > (larger than symbol). Here is an example of the .txt file:

>>128

10347

00>17:05:38,06/27/12,78.0,920

01>17:05:42,06/27/12,78.0,920

02>17:05:45,06/27/12,78.0,920

I would like the outcome to look something like this:

00 17:05:38 06/27/12 78.0 920 01 17:05:42 06/27/12 78.0 920 02 !7:05:45 06/27/12 78.0 920

Or even better, if I could ignore the first column and show the time and date columns as serial time and date. So have only 4 columns like this:

0.71224537 39625 78.00 920 ....

The last step could probably be done after the import.

So far I have been trying to use this code but without much luck :

fd = fopen('file.txt','rt');
data = textscan(fd, '%d%d%f%d', 'delimiter', ', >','headerlines',5);
fclose(fd);

I hope I described my problem good enough and someone will be able to help me. Any hint into the right direction is very appreciated...

Thank you very much...

Nils

Answer 1

[Walter Roberson]

dataCell = textscan( fid, '%*3c%s%s%f%f', 'Delimiter', ',', 'HeaderLines', 2 );

Then

sernums = datestr( strcat(DataCell(2), ' ', DataCell(1)) );
output = [sernums dataCell{3} dataCell{4}];

This would convert the times and dates together as serial date numbers. If you want the columns to be separate as you show in your example, then

sernums = [datestr( DataCell(2) ), datestr( DataCell(1) )];
output = [sernums dataCell{3} dataCell{4}];

[Nils]

Hello Walter,

thank you for your fast response. The import seems to work now, but I get an error with the time/date conversion. This is the code:

fid = fopen('data.txt','rt');

dataCell = textscan( fid, '%*3c%s%s%f%f', 'Delimiter', ',', 'HeaderLines', 4 );

sernums = [datestr( dataCell(1,2) ), datestr( dataCell(1,1) )];

output = [sernums dataCell{3} dataCell{4}];

... and this is the error:

??? Error using ==> datestr at 180 Cannot convert input into specified date string. DATENUM failed..

Also, could you explain a little bit the textscan code so I could change it if needed? What does "%*3c" do? And is it possible to obtain a double rather than a cell matrix? I hope these questions are not stupid.

Thank you so much.

Nils