Import large csv as datastore with formatting errors

Hi,

I am trying to import a very large CSV file as a tall table. Right now I am just testing the demonstration code, I have run into a problem! At least one row in my CSV file has a formatting error. The error prevents any code from working.

ds = tabularTextDatastore('data.csv');
tt = tall(ds);
ids = unique(tt{:,25});
ids = gather(ids);

This code produces the error.

Evaluating tall expression using the Local MATLAB Session:
- Pass 1 of 1: 99% complete
Evaluation 90% complete
Error using matlab.io.datastore.TabularTextDatastore/readData (line 77)
Mismatch between file and format character vector.
Trouble reading 'Numeric' field from file (row number 17128, field number 50) ==> "    .","",""," .","   .",".",132400,"-1.0","X",2008\n
Learn more about errors encountered during GATHER.

Error in matlab.io.datastore.TabularDatastore/read (line 120)
            [t, info] = ds.readData();

Error in tall/gather (line 50)
[varargout{:}, readFailureSummary] = iGather(varargin{:});

Caused by:
    Reading the variable name 'Var50' using format '%f' from file: 'data.csv' starting at offset 2583691484.

I know from previous experience with this data set that one of the lines had a formatting error that then causes more errors. The file is too large to open and edit manually. Is there anyway for me to fix this error in matlab or throw out the bad lines?

Thanks!

Answer 1

[per isakson]

Comment out the offending rows

open files
while not( feof(fid) )
    chr = fgetl(fid);
    is_ok = analyse_row( chr );
    if is_ok
        fprintf( fid_out, '%s\n', chr )
    else
        fprintf( fid_out, '%% %s\n', chr )
    end
end
close files

Test with small files. Five good rows and a couple of faulty rows.

Read the new file with CommentStyle,'%'

[Russell Shomberg]

This is great! I can also write a file with all the bad rows!