textscan of csv files produces empty rows and cols

I have been using textscan to read large batches of CSV files, and every once in a while, I'd get a file that has "invisible" blank lines that get read into my data cellarray. Sometimes its blank rows, sometimes its blank columns, but they occur at end of the data only.

I'm trying to remove those blank rows and cols.

I'm using:

data(cellfun(@(X) isempty(X), data(:,1)), :) = [];

but its not working. I debugged it, and apparently matlab doesn't think those empty rows are empty.

Running isempty on one of those empty rows yeilds a value of 0, as wel as running on a row that DOES contain data

Anyone experience the same thing?

[Walter Roberson]

Could you show the content of one of those rows? Also, size(data{K,1}) and class(date{K,1}) where K is one of the rows that appears empty?

[David C]

size was 1 and 1, class was cell

[David C]

the answers are the same if i test the cells that actually contain values too.

this problem goes away if I physically got into the CSV file with excel and delete those rows.

super strange

[Walter Roberson]

Okay, and to double-check, you used

size round-bracket data curly-bracket K comma 1 curly-bracket round-bracket

?

If so then your cell array location is not empty because it contains a cell array (which might be empty)

Maybe,

data(cellfun(@(X) isempty(X) || (iscell(X) && all(cellfun(@(Y) isempty(Y),X)), data(:,1)), :) = [];