textscan not eating EOL

Hi.

I'm trying to read the following (tab separated values) file with textscan :

-------------------------------------

year  day in year  day in month  month  hour
Y  DY  DM  M  H
1900  1  1  1  1
1900  1  1  1  2

-------------------------------------

I would use the following code :

-------------------------------------

fid=fopen('test.txt', 'rt');
nb_col = 5;
t = textscan(fid, repmat('%s ', 1, nb_col),1, 'Delimiter','\t\n', 'CollectOutput',1);
headers = t{1};

t = textscan(fid, repmat('%f ', 1, nb_col), 'Delimiter','\t', 'HeaderLines', 1, 'CollectOutput',1);
values = t{1};
fclose(fid);

disp(headers)
disp(values)

-------------------------------------

Except this doesn't work. It seems the EOL of the header line is not eaten by textscan after the first call, and HeaderLines counts from there in the second call.

I need to enter 2 as HeaderLines.

t = textscan(fid, repmat('%f ', 1, nb_col), 'Delimiter','\t', 'HeaderLines', 2, 'CollectOutput',1);

Another workaround I found here is to enter \t\n as delimiter, although I don't really understand it since my delimiter is only \t and \n is only present at the end of the line.

t = textscan(fid, repmat('%s ', 1, nb_col),1, 'Delimiter','\t\n', 'CollectOutput',1);

t = textscan(fid, repmat('%f ', 1, nb_col), 'Delimiter','\t\n', 'HeaderLines', 1, 'CollectOutput',1);

Is this a known issue ?

I'd call it a bug, since textscan does not show the behavior I expect from it reading the doc, but it might be my understanding that is incorrect.

Thank you for your help.

Answer 1

[Walter Roberson]

Perhaps

'Delimiter',{'\t', \n'}