Thread Subject: %s vs. %c in textscan with multiple delimiters

I've run into a problem with using textscan to parse lines with multiple null fields. For example:

>> line='a,b,c,d,e,f,g';
>> foo=textscan(line,'%s','delimiter',','); foo{:}'
ans =
    'a' 'b' 'c' 'd' 'e' 'f' 'g'

>> foo=textscan(line,'%c','delimiter',','); foo{:}'
ans =
abcdefg

... works as expected. However,

>> line='a,b,c,d,,,g';
>> foo = textscan(line,'%s','delimiter',','); foo{:}'
ans =
    'a' 'b' 'c' 'd' '' '' 'g'

>> foo = textscan(line,'%c','delimiter',','); foo{:}'
ans =
abcd,g

... so when scanning with %s, the "identify multiple consecutive delimiters" functionality works, but with %c it fails. Is this expected behaviour, or is this a bug?

Thanks in advance for any guidance.

/HJ