Regexp dynamically naming tokens issue

I'm trying to parse some text using named tokens, but I've run into an issue.

Running the following code does exactly what I expect.

test = '<Field name="test1">data1<Field name="test2">data2</Field>';  
format = '<Field name="([^"]+)">(?<test1>[^<>]*)</Field>';  
output = regexp(test,format,'names')

output = 
    test1: 'data2'

Now I'd like to have it collect all fields with a single regular expression. I tried many different formats (one example is replacing (?<test1>[^<>]*) with ((?<$1>[^<>]*)) ), but I'm working with the following since I thought it would be easy to diagnose: (??@cat(2,''(?<'',$1,''>[^<>]*)''))

test = '<Field name="test1">data1<Field name="test2">data2</Field>';
format = '<Field name="([^"]+)">(??@cat(2,''(?<'',$1,''>[^<>]*)''))</Field>';
output = regexp(test,format,'names')

output =   
0x0 struct array with no fields.

I've been assuming that I was just generating the string incorrectly, but outputting to console results in:

test = '<Field name="test1">data1<Field name="test2">data2</Field>';
format = '<Field name="([^"]+)">(?@cat(2,''(?<'',$1,''>[^<>]*)''))</Field>';
output = regexp(test,format,'names')

ans =

(?<test1>[^<>]*)

ans =

(?<test2>[^<>]*)

output = 

0x0 struct array with no fields.

I'm generating what I intended with the dynamic expression, but something obviously isn't working. It's odd that replacing the dynamic portion of my regular expression with what it evaluates to works, but the dynamic expression doesn't work.

Any help would be appreciated. I'm running MATLAB R2015b if it helps.

Thanks, Garrett

Answer 1

[Stephen23]

It might be possible to solve this using dynamic regular expressions, but I think in the interest of code clarity it would be best to solve this task outside of regexp. Doing so will make the code intent clearer, make debugging easier, and probably run a bit faster (dynamic expressions get evaluated).

Something like this:

test = '<Field name="test1">data1<Field name="test2">data2</Field>';
fmt = '<Field name="(\w+)">([^<>]*)';  
C = regexp(test,fmt,'tokens');
C = vertcat(C{:})';
S = struct(C{:})

gives

>> S
S = 
    test1: 'data1'
    test2: 'data2'

Tip

Developing regular expressions can be a challenge, you might like to try my Regular Expression Helper tool, which is an interactive figure that parses a regular expression as you type and gives immediate feedback on regexp's outputs:

http://www.mathworks.com/matlabcentral/fileexchange/48930-regular-expression-helper

[Garrett]

Interesting, I wouldn't have thought to do it this way, but it works! Thanks :)