How does one input a large amount of data that has a different number of columns per row?

Hi, I am trying to load in a few thousand files using textscan that are in the following format:

3245

0,1,2,3,4,......500

0.23,0.35,0.41,0.57,1.32

0.10,0.83,0.92,1.45

0.66,0.94,1.75,2.35,2.51,3.90

0.13,0.71

I only want to load starting at the 3rd row (i.e. at 0.23). However, starting at the third row, there are actually 500 more rows and a variable number of columns (on the order of a few hundred). (I don't need to load in the first two lines). But I can't figure out how to do this. I can't input them row by row, because there are 500 of them, and a few thousand files. I want to be able to access each number separately of course (say data(2,3)=0.92), and to fill in the extra spaces with NaN or something. Can someone please help me with this?

Kyla

Answer 1

[per isakson]

textscan cannot parse rows with variable number of columns. AFAIK: there is no function in Matlab that can do that.

There are many approaches to read the file. Two of them:

1. one line at a time with fgetl
2. read each line as a string with textscan and parse in a second step

.

--- textscan CAN parse rows with variable number of columns ---

I was wrong. This code parses the first NCOL columns, where NCOL is the number of columns of the shortest row.

    fid = fopen('h:\m\cssm\ascii_with_variable_row_length.txt');
    cac = textscan( fid, '%f%f%*[^\n]', 'CollectOutput', true );
    fclose( fid );
    cac{1}

prints

    ans =
        17    24
        23     5
         4     6
        10    12
        11    18

where 'h:\m\cssm\ascii_with_variable_row_length.txt' contains

    17    24     1     8    15
    23     5     7     
     4     6    13    20    22
    10    12     
    11    18    25     2     9

[Kyla]

OK, thank you for letting me know!

[per isakson]

Sorry, now I know better! See above.