How can I import a .txt file correctly using textscan?

I am having a horrible time trying to import this data set into matlab:

http://www.cftc.gov/files/dea/history/com_disagg_txt_2012.zip

It's a bunch of agricultural futures information, and so it has text as well as numerical data. I would like to import it so that I can read and utilize the dates and data (would like to have the text, but not as important). Anyway, when I use textscan I get ridiculous results and they're unusable.

for example:

fid = fopen('c_year.txt');

data = textscan(fid,'%s %d %s %d ...and so forth for 188 columns', 'delimiter', ',','headerlines',1);

fclose(fid)

I need it not to import as a struct or cell array, I need to be able to split that up afterwards.

Thanks

[Walter Roberson]

textscan() always creates a cell array.

Answer 1

[Geoff]

The way to have the most control over a line-based txt file import is to read it one line at a time:

while !feof(fid)
  linestr = fgetl(fid);
end

Then use textscan or sscanf on the line after reading it.

doc sscanf

If you prefer, you could use regexp to split out the fields from the comma-separated list.

% Returns a cell array of each field (empty fields handled)
toks = regexp( linestr, '([^,]*),|([^,]*)$', 'tokens' );

[Mango]

I'm not sure if you took a peek at that file, but it essentially has one header line, the first column is all text, then the rest are mixed and matched, some text, some integers, some decimals, etc. I'm not great with MATLAB, so I'm not 100% sure on how you're saying for me to go about it.

so I would write what?

fid = fopen('c_year.txt'); while !feof(fid) linestr = fget1(fid); end

doc textscan(????)

then it will export to a cell array of each field?

Thanks so much for helping.