Alias blanks in the cell array as a string.

I am using the ‘importdata’ function to read a text file. It reads the numeric data as a matrix and text data as a cell array; now the first element of cell array is all the headers, I use a ‘strsplit’ function to separate each individual variable name in header. The issue is that there are some blank columns in the text file, so the numeric matrix reads it as a ‘NaN’ but the resulting variables (which I used ‘strsplit’ for) eliminates blanks. Now the result is that size(textdata,2) +size(data,2) is not equal to size(variable,2). I want to read blanks in the header as a variable, alias it as ‘empty’

5 Comments

"I am using the ‘importdata’ function to read a text file..."
There's the root problem. importdata, while handy from a simple syntax standpoint tends to create more trouble than it solves as you discovered. I'd recommend using readtable instead and you'll have both variables handled by name by their proper type in a single table variable. You can customize how things are imported regarding the missing columns per your requirements with either some additional named variable options or by the use of an import object via detectImportOptions
I second dpb's comment: avoid IMPORTDATA, use READTABLE or TEXTSCAN or the like.
Thank you all. I will experiment with 'readtable' and 'textscan' some more. Really appreciate this community for encourging the correct way of programming.

Sign in to comment.

 Accepted Answer

As the OP found, the solution is to include multiple delimiters
strsplit(t,'\t', 'CollapseDelimiters',false)
And as others have recommended, using readtable offers better control of how data are read than importdata.

More Answers (1)

Hello All, taking advice of everyone who commented; I was able to robustly solve the issue of dowloading data from a text file and also retain the variable name. This particular text file had variable names in line 1 and line line 1 also had blanks. Following code has worked reliably for me.
And if your text file happen to have many lines before the line with headers, following has worked well for me.

Categories

Products

Release

R2015a

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!