reading whitespace in textscan

Hi,

I have some data to read in fixed field format as date/time (ddmmyyyyHHMM) then 4 data fields, as per this example:

24 91987 530    0.8455    0.7200   -0.1255   -0.1255
30 919872345    0.8329    0.7500   -0.0829   -0.0829
 1101987 0 0    0.8687    0.8000   -0.0687   -0.0687

My code is:

       resarr=textscan(fidres,'%2f%2f%4f%2f%2f%10.4f%10.4f%10.4f10.4f','whitespace','','delimiter','/n');
  res= [datenum([resarr{:,3},resarr{:,2},resarr{:,1},resarr{:,4},resarr{:,5},zeros(size(resarr{:,1},1),1)]),resarr{:,6},resarr{:,7},resarr{:,8},resarr{:,9}];

I'd like to read the data and convert the date to datenum. But am having a lot of trouble with the whitespaces.

eg the first line gives   24,91,987,53,0,0.8455,0.7200,-0.1255,-0.1255
instead of                24,9,1987,5,30,0.8455,0.7200,-0.1255,-0.1255

How should I change the textscan?

Answer 1

[Hugo]

I can only think of two options:

1) Add spaces separating the fields in the data, i.e. instead of ddmmyyyyHHMM, modify the data so that you have dd mm yyyy HH MM. That should be easy.

2) The other way is to scan for text

resarr=textscan(fidres,'%2s%2s%4s%2s%2s%10.4f%10.4f%10.4f10.4f','whitespace','','delimiter','/n');

and then convert the text to numbers using, for example, str2double or str2num