problem in using textscan to import data from csv

I need to import a csv file with 60 columns. All have numerical values. However, some values are missing and instead of keeping it as 0 or Nan, those columns have "*****" and when importing using textscan it gives an error.

So how can I replace those string parts (******) with NaN and import it as numbers.

so finally my requirement is to import the csv file to number arrays with replacing "*****"with NaN

[Walter Roberson]

Could you give a couple of sample lines? And also please show your current textscan attempt.

Answer 1

[dpb]

Filling a field with asterisks is Fortran convention if output magnitude exceeds available width for a field; possibly can you rework the application building the output file to correct such?

Failing that, use the 'TreatAsEmpty','******' option to cause textscan to return those fields as missing values.