How to find the end of a word using strpos()

Question

I'm getting an address input as one long string as opposed to name, address, city, state and zip. I split most of it up, besides for between the address and the city. So I want to search for any street type name (court, road, street, avenue, etc) and then split the string at the end of the word. Then I will be left with the address and city separate.

strpos() only gives me the position of the beginning of the keyword, I want it to split at the end of the keyword. For example this is where I'm up to:

John Doe
1 Main Street Anytown
NY
00000

I want to split between Street and Anytown. And this address won't be static, there may be more words etc.

Another idea would be a function that automatically splits a string into different fields. Someone told me that in some countries the postal service has an API that does it. Does USPS have such a thing? Their site doesn't indicate it.

USPS has API to check weather it is valid address , if you gave an address in correct format. But that doesnot solve the problem here — mujuonly, Jul 28 '16 at 08:09
You mean strrpos function? (strrpos — Find the position of the last occurrence of a substring in a string) — Maxim Tkach, Jul 28 '16 at 08:25
Once you get beginning of keyword with `strpos`, you can just add `strlen("keyword")` to get end (specifically the space after the keyword). — Markus Laire, Jul 28 '16 at 08:26
@MaximTkach no, i want to find the "t" in street, or the "e" in avenue etc... The last letter position in the string — Mennyg, Jul 28 '16 at 08:27
seems answered here http://stackoverflow.com/a/3923215/6394185 — Mehmet SÖĞÜNMEZ, Jul 28 '16 at 08:36
FWIW, if you're fighting on the level of `strpos`, you're probably gonna lose the larger battle against this problem. You seem to be making very naïve assumptions about what a street address will look like and how it can be split from its city name. That *may* work 80%, 90% of the time, but you *will* fall into false positives with this simple approach. Typically you need an actual address database to resolve against and find the best canonical match from. — deceze, Jul 28 '16 at 08:42
@Mujeebu mentioned the USPS API; unless you're prepared to dive really deep into this topic, I'd suggest an existing API as well (not that I can make any specific recommendations). — deceze, Jul 28 '16 at 08:46
A more important question is what are you trying to achieve with the address? Even address resolution approaches are still problematic and I don't really have any issues with you using a Bigdata approach. If the core of the problem is address validation, then you should definitely use a library. If however, you're just trying to extrapolate information for storage/matching, then you can go ahead with your current implementation. — Chibueze Opata, Jul 28 '16 at 08:47
@deceze i think I found something that'll work for me... http://stackoverflow.com/a/10695075/6442152 — Mennyg, Jul 28 '16 at 08:51

score 1 · Answer 1 · answered Jul 28 '16 at 08:39

You could do something like this.

$word = "Street"

$pos = strpos($word, $address);
if ($pos !== false) {
    $pos += strlen($word) - 1;
    ...
}

Depending on how many different words to match there are, using regex might work better than using string functions.

score 1 · Answer 2 · answered Jul 28 '16 at 08:59

Actually what you should do is searching for multiple needle in haystack, summing position and length of needle.

function strpos_street($string){
    $types = [
   'court', 
   'road', 
   'street', 
   'avenue' 
  ];
  foreach($types as $t){
      $pos = strpos(strtolower($string), strtolower($t));
      if($pos === false) continue;
      else return $pos + strlen($t) - 1;
  }
  return -1;
}

echo strpos_street($string);

Some Tips:

Consider case sensivity

You may use explode function in a same iterative way

I'm not sure but there may be some addresses which contains both "street" and "avenue" words

Chibueze Opata · Answer 3 · 2016-07-28T08:51:14.773

It seems that you're trying to search in reverse as strpos gives position of a string starting from the beginning by default, you will need to set the offset to -1 and subtract from strlen.

i.e.

strlen($address) - strpos($address, 'Anytown', -1) - 1;
//search from reverse and output position starting from reverse

However as pointed out in the comments, this may not the best approach for the problem you're trying to sovle.

score 0 · Answer 4 · edited Jul 28 '16 at 09:14

0

You can use explode function, which split a string using keyword and return an array of strings. Example of using explode:

$str = '1 Main Street Anytown';        
$array_of_string = explode('street',$str);

Then $array_of_strings[0] will contains 1 Main and $array_of_strings[1] will contains Anytown

edited Jul 28 '16 at 09:14

R. Zagórski

20,020
5
65
90

answered Jul 28 '16 at 08:51

A.Tarhini

72
4

How to find the end of a word using strpos()

4 Answers4