[R] named character question

[David Winsemius]

On Aug 12, 2012, at 8:33 PM, Erin Hodgess wrote:

> Dear R People:
>
> Here is a goofy question:
>
> I want to extract the zip code from an address and here is my work  
> so far:
>
>> add1
>                  results.formatted_address
> "200 W Rosamond St, Houston, TX 77076, USA"
>> add1[1][32:36]
> <NA> <NA> <NA> <NA> <NA>
>  NA   NA   NA   NA   NA
>> str(add1)
> Named chr "200 W Rosamond St, Houston, TX 77076, USA"
> - attr(*, "names")= chr "results.formatted_address"

 > ttt <- "200 W Rosamond St, Houston, TX 77076, USA"

 > sub("^.+,.+,\\s[[:alpha:]]*\\s([[:digit:]]{5}).+", "\\1", ttt)
[1] "77076"

You will need to determine if all you addresses have two commas before  
the two letter state designation. You may not need as specific a  
pattern as this. An alternate pattern.

 > sub("^.+\\s[[:alpha:]]{2}\\s([[:digit:]]{5}).+", "\\1", ttt)
[1] "77076"

-- 

David Winsemius, MD
Alameda, CA, USA