Importdata won't work with this txt file

Hello, I want to take a txt file that contains headerlines and then a matrix of numbers that represents an image. I cannot get importdata to read the file, I get

??? Error using ==> importdata at 214 Unable to load file. Use TEXTSCAN or FREAD for more complex formats.

I also can't get textscan to work. I have attached the file

//-----------------------------------Begining of file----------

File : C:\
Title: 
Date : 17/06/2011

Detector 31, NSCG Surface 1: 1mm sample
Size 1.000 W X 1.000 H Millimeters, Pixels 200 W X 200 H, Total Hits = 158280
           1               2                  3
   1  9.781E-003  1.676E-002  2.619E-002
   2  2.224E-002  2.936E-002  3.852E-002
//--------------------END OF File

note, there are tabs at the begining of lines and there are 200 rows and columns of data. I dont need to read in the column or row numbers

[Jason]

I've just checked and my program that creates these txt files was set with the txf file encoding as "Unicode". When I switch it to ANSI, import data and textscan work. The problem is that I have all my data in this UNICODE format. Can importdata or textscan work with this?

Answer 1

[Gerd]

Hi Jason,

I would use the textscan command to extract the data. By using

fid = fopen('Data.txt')
tmp = textscan(fid,'%s%f%f%f')
fclose(fid)

you will get a cell array of data. Now you can extract the data row by row using

cell2mat(tmp(index))

Now you are able to extract whatever you want out of your data. I cannot test this script in detail but I am pretty sure with some examinations on the data it should work.

If you do have problem please post.

Gerd

[Gerd]

You could also use fgetl(fid) to examine the data row by row

[Jason]

Hu, Im sorry but I still can't get it working with textscan. When I use your second suggestion fgetl, then the line that is read in has for some reason put an extra space between the characters!

>> tmp=fgetl(fid)

tmp =

F i l e : C :

[Gerd]

Jason,

what exately do you want to extract from the file?

If you use

tmp=fgetl(fid)

you get your data line by line.

e.g. if you want to seperate the first row 1 9.781E-003 1.676E-002 2.619E-002

you have to use

tmp_split = textscan(tmp,'%f','delimiter','\t')

Now you have a cell array with 4 double values. The first one is your row which you don't need and the rest is your data.

[Jason]

hi Gerd, I would like to be able to pull out a single matrix and thats why I thought textscan would do thr trick. Whilst fgetl works, it doesnt feel as elegant as I need to work out the number of times to perform fgetl, and then execute this that many times.

[Jason]

I've also noticed when I use the fgetl(fid) way, there are strange symbols at the very begining that are't present when I open in notepad. Could this be the cause of why importdata and textscan don't work?

>> tmp=fgetl(fid)

tmp =

ÿþ 1 6 . 1 8 4 E - 0 0 1 6 . 1 8 4 E - 0 0 1 6 . 1 7 0 E - 0 0 1

>> fclose(fid)

[Gerd]

I don't see the problem with my file. Maybe it is possible for you to put the file somewhere on the net

[Jason]

Could I email it to you? as i don't see any upload facility here.

[Gerd]

Sure you can do but I will be out of office until tomorrow.

bwts.tre at googlemail dot com