textscan gives me wrong number of lines

I have a file names inputR_revised.tsv at https://www.dropbox.com/s/vtby4027rvprhga/inputR_revised.tsv?dl=0 In matlab, I typed

    fid=fopen('BMC3C/example/inputR_revised.tsv','r')
    covTable = textscan(fid,['%s',repmat('%.8n',[1,20])],'HeaderLines',1);

I get covTable{1,1} of size 41699 times 1. However when I type the following at terminal

    wc -l inputR_revised.tsv

I get 41677. Why does it differ? I have used sed and cut to modify the original file to get inputR_revised.tsv. Is this the reason?

Is there a way to fix this?

Answer 1

[dpb]

In

covTable = textscan(fid,['%s',repmat('%.8n',[1,20])],'HeaderLines',1);

you set the 'HeaderLines' option to True (1) so the first line is being ignored by textscan and read as a header, not data. If it is actually data instead, then

covTable = textscan(fid,['%s',repmat('%.8n',[1,20])],'HeaderLines',0);

will fix that (or error if it really is a header, one; we can't tell whether it is or isn't).

[Sumin Kim]

Thank you. But there is an header. And the number of lines differ by 22

[dpb]

Attach the file here...