How to import csv files with more than 100000 columns using csvread

I'm trying to import csv files that have 25 rows and >100000 columns of data points.

When the file contains more than 100000 columns or data points per row, csvread is importing the data as a vector rather than matrix. This code reproduces the error on my machine:

x = rand(25,110000);

csvwrite('test.csv',x)

test = csvread('test.csv');

I would like test to be the same size as x but I get vector of size 2750000x1.

Any ideas?

[Jeremy Hughes]

Hi Sean,

CSVREAD stops trying to count the number of columns at 100,000 and assumes the file contains just one long vector.

I suggest using TEXTSCAN. This code should do what you need.

fid = fopen('test.csv','r');
fmt = repmat('%f',1,110000);
d = textscan(fid,fmt,'Delimiter',',','CollectOutput',true);
test = d{:};
fclose(fid);

Hope this helps,

Jeremy