You are welcome. Glad you liked it.

Cool!

I knew there was a nice way to do it with replacement, but I was torn because I wanted to focus first on regexp. Thanks for sharing this – because ultimately it is probably the most efficent and cleanest way to do it.

Maybe those flashcards would be good to share on File Exchange. A cheat sheet for people to hang in their offices might be really nice.

regexprep(locationNames,’([A-Z].)[^A-Z]*’,'$1′)

:)

I chatted with development, and let them know of your request. However, right now I think your way of doing it as “easy” as it can get. There isn’t anything that you are missing that I am aware of. I will try and think about it some more and see if anything else pops into my brain.

Thanks,
Sarah

>> strCell = {‘first’,'second’,'third’,'fourth’};

Using strcmp gives me a nice little logical mask that matches my criteria:

>> strcmp(‘third’,strCell) % [0 0 1 0]

Let’s say I had some criteria that needed regular expressions, and wanted to get a logical mask telling me which cells matched my criteria.

>> regexp(strCell,’(f|r)’)

ans =

[1x2 double] [] [4] [1x2 double]

That, of course can be simplified a bit by using ‘once’:

>> regexp(strCell,’(f|r)’,'once’)

ans =

[1] [] [4] [1]

But it’s still not a logical mask. The only way I can think to make it one is to wrap it in a cellfun like this:

>> ~cellfun(@isempty, regexp(strCell,’(f|r)’))

ans =

1 0 1 1

Now, the above of course works just fine, but it just feels a touch awkward… Asking “which strings match my pattern” seems like a very simple (and very often asked!) question, but getting an answer to it involves the negation of a cellfun() call.

It feels like strcmp() is nice and straight forward, but adding the (relatively minor) addition of matching a pattern instead of an exact string takes a detour into logic and functions that makes the (notoriously confusing) regular expressions that much more tricky in MATLAB.

Thoughts? Have I simply missed an obvious regexp() parameter that would make it behave how I want it to?