Why textscan inserts 0 instead of NaN ?

Hello,

I use textscan to import a text file which is a mixture of number and strings. The format is like:

VGAHFID,VGALFID,SN_hd

63484,49736,S410060Y_0

54696,54744,S410060Y_0

,,S410060Y_0

53092,54512,S410060Y_0

53716,54568,S410060Y_0

53800,54768,S410060Y_0

The command I used is a = textscan(fid,'%d%d%q','Headerlines',1,'Delimiter',',','EmptyValue',NaN);

Here's the result.

>> a{2} ans =

       49736
       54744
           0
       54512
       54568
       54768

If I changed the NaN to Inf, the missing value will be replaced with 2147483647.

I guess it must be my "a" format issue. But don't know how to solve it.

Thanks, Jane

[Jane]

Never mind. I checked NaN, it is double. So once I changed the format to %f, the NaN showed up.