Match each address from the address number to the 'street type'

Question

I have a paragraph of text that contains the following addresses:

at 900 Greenwood St.
in 500 block of Main Street
at 670 W. Townline Ave.
before 1234 River Avenue
of 1125 Main Ave.

I want to match each address from the address number to the 'street type' (ave., street, lane, road, rd., etc.) except for addresses that begin with the word of.

So of the addresses above, the statement would match:

900 Greenwood St. 500 block of Main Street 670 W. Townline Ave. 1234 River Avenue

and would not match:

1125 Main Ave.

Will the paragraph ever have numbers that are not part of an address? — ridgerunner, Jul 01 '11 at 00:51
Even after applying one of the regex's below, you'll want to confirm that it was parsed correctly. I work in the address validation industry (SmartyStreets) where this is a common task -- so to make sure you have accurate data, you'll want to test it against USPS-recognized addresses. For example, look into LiveAddress (it's free). Hope that helps. — Matt, Jan 19 '12 at 19:54

score 2 · Answer 1 · answered Jul 01 '11 at 00:29

As far as I know, there isn't a one simple regex pattern for this kind of complicated task. There are too many variables to cover for one pattern to work reliably. My first guess would be to look for "street", "ave", etc., but what if the street name doesn't have a suffix (i.e. 999 La Canada)? You could look for any phrase between "at", "in" or "before", but what if one of those phrases isn't an address? See what I mean?

My suggestion would be to take a look at Lingua::EN::AddressParse for Perl.

Mike Samuel · Answer 2 · 2011-07-01T06:13:21.943

1

When

s = "at 900 Greenwood St.\n\
in 500 block of Main Street\n\
at 670 W. Townline Ave.\n\
before 1234 River Avenue\n\
of 1125 Main Ave."

the regex

/(?:^|\s)(?:(?!of\b)[a-z]+)\s*(\d[\s\S]*?\b(?:ave\.|avenue|st\.|street|lane|road|rd\.))/gi

used thus

var addresses = [];
for (var match = [], re = /(?:^|\s)(?:(?!of\b)[a-z]+)\s*(\d[\s\S]*?\b(?:ave\.|avenue|st\.|street|lane|road|rd\.))/gi;
     match = re.exec(s);) {
  addresses.push(match[1]);
}

produces

["900 Greenwood St.","500 block of Main Street","670 W. Townline Ave.","1234 River Avenue"]

edited Jul 01 '11 at 06:13

answered Jul 01 '11 at 02:48

Mike Samuel

118,113
30
216
245

@Brock Adams, \w matches [0-9] which made it accept partial numbers. That is now fixed. – Mike Samuel Jul 01 '11 at 06:14

score 1 · Answer 3 · answered Jul 01 '11 at 05:59

This is fulfilling your request:

(?!^of\b)^.*?(\d+.*?(?:St\.|Street|Ave\.|Avenue))$

See it here on Regexr

(?!^of\b) Negative look ahead, row does not start with the word "of"

^ Matches the start of a row, use the m modifier!

.*? matches everything non greedy

(\d+.*? when the first numbers are found start the first capturing group with the (

(?:St\.|Street|Ave\.|Avenue)) Non capturing group because of the ?: matches the alternations between the |. The last ) closes the capturing group with the result.

$ Matches the end of the row, use the m modifier!

Your result is in the first capturing group.

Important this is working with your given examples, addresses can be that different, it will not work on all kind of existing addresses.

score 0 · Answer 4 · answered Jul 01 '11 at 00:18

var addrs = create_array_of_possible_addresses();
var matching_addrs = [];
for (var i=0; i < addrs.length; i++) {
    if ( addrs[i].match(/^of/) continue;
    if ( addrs[i].match((/\d.*(?:St\.?|Street|Ave\.?|Avenue|Ln\.?|Rd\.?|Road))/ )
        matching_addrs.push( RegExp.$1 );
}

Untested.

Match each address from the address number to the 'street type'

4 Answers4

Linked