We can start…

The sample is going to work with this table:

We will create the procedure called gpsdistance with two input numbers mybuilding_id and mydistance.

mybuilding_id represents ID number of our base building in which surroundings we will search further buildings (which are situated inside the distance handed in the second parameter mydistance) (in kilometers!).

Sequentially we will declare these internal variables:

• mylon – longitude of our base building
• mylat – latitude of our base building
• lon1 – 1. longitude limit
• lon2 – 2. longitude limit
• lat1 – 1. latitude limit
• lat2 – 2. latitude limit

Next command loads actual values of our base building into variables mylon and mylat:

Follow calculating of limiting coordinates on the basis of mydistance. Number 111 represents length of 1° of latitude in kilometres (69 miles). The meaning of formulas you can find in the original document (link below).

And engine of the whole sample is a select which searches buildings located from the base building up to the value of mydistance. In addition this script returns column called distance which represent real distance (in kilometres) from the base building. Number 6372.795 is earth radius in kilometres (3956 miles). Again, the meaning of formulas you can find in the original document (link below).

Complete MySQL procedure

Command DELIMITER // set separator on two slashes. It’s happening because the default separator is semicolon. Semicolons are emerged inside the procedure and it would cause the fault during procedure creation. That’s why you should pay attention as long as you create procedure in phpMyAdmin (you will have to set separator different from semicolon – under the SQL command field).

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
DELIMITER //

CREATE PROCEDURE `gpsdistance` (IN mybuilding_id int, IN mydistance int)

BEGIN

declare mylon double;

declare mylat double;

declare lon1 float;

declare lon2 float;

declare lat1 float;

declare lat2 float;



SELECT building_latitude, building_longitude INTO mylat, mylon FROM buildings WHERE building_id = mybuilding_id;



SET lon1 = mylon-mydistance/abs(cos(radians(mylat))*111);

SET lon2 = mylon+mydistance/abs(cos(radians(mylat))*111);

SET lat1 = mylat-(mydistance/111);

SET lat2 = mylat+(mydistance/111);



SELECT buildings.*, 6372.795 * 2 * ASIN(SQRT(POWER(SIN((mylat-ABS(building_latitude))*pi()/180/2), 2) +

COS(mylat * pi()/180) * COS(ABS(building_latitude) * pi()/180) * POWER(SIN((mylon - building_longitude) * pi()/180 / 2), 2))) AS distance

FROM buildings WHERE building_latitude IS NOT NULL AND building_longitude IS NOT NULL AND

building_longitude BETWEEN lon1 AND lon2 AND building_latitude BETWEEN lat1 AND lat2 HAVING distance < mydistance ORDER BY distance LIMIT 50;



END//

Calling the procedure in PHP

Very important is to know that you can call procedures only through MySQLi layer (which is object-oriented). The procedure can be call by this SQL command: call gpsdistance($mybuilding_id, $mydistance). Here is a short sample:

1
2
3
4
5
6
7
8
9
10
11
12
$link = mysqli_connect('localhost', 'db', 'user', 'passwd');

if (!$link) {

    printf("Can't connect to MySQL Server. Errorcode: %s\n", mysqli_connect_error());

    exit;

}

if ($result = mysqli_query($link, "call gpsdistance($mybuilding_id, 5)")) {

    while($row = mysqli_fetch_array($result)) {

        ...

    }

    mysqli_free_result($result);

    mysqli_close($link);

}

That’s all. I hope the article is useful for somebody

Used literature

• RUBIN, Alexander. Geo/Spatial Search with MySQL. Scribd [online]. 2008 [cit. 2009-11-04]. Accessible from WWW: <http://www.scribd.com/doc/2569355/Geo-Distance-Search-with-MySQL>.

The geocoding service may only be used in conjunction with displaying results on a Google map.

Geocoding Service

Geocoding Service provides way to get geographic coordinates via an HTTP request. We must send a request to http://maps.google.com/maps/geo? with several parameteres. We are interested in these parameters:

• q (required) – the address to which you want get coordinates
• key (required) – your API key (you can obtain it here)
• sensor (required) – true if the request is sent from a device with location sensor, otherwise false
• output (required) – the format of output (the options are xml, kml, csv or json)
• oe (optional) – output encoding (It’s recommended set this parameter to utf8)
• …a další

The input data should be encoded in utf-8.

CSV output format

Output data can be generated in these formats:

• json (default)
• kml
• xml
• csv – data are separated by semi-colon

In our situation is CSV ideal because it returns only 4 blocks of data separated by semi-colon:

1. HTTP status code
   • 200 - succes
   • 500 – server error
   • 602 – unknown address
   • 610 – bad API key
   • 620 – too many queries (>5000 per day or too many requests in too short a period of time)
   • a další…
2. Accuracy
   • 0 – unknown
   • 1 – country level
   • …
   • 4 – city (village) leel
   • 5 – ZIP level
   • 6 -street level
   • 7 – intersection level
   • 8 – address level
   • 9 – building level
3. Latitude
4. Longitude

For example:

PHP: example of getting GPS coordinates from CSV file

We assume having table in db with buildings’ addresses to which we can find their GPS coordinates. The table could look like this:

Then our PHP script for filling up GPS coordinates could look like this:

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
<?php

$conn = mysql_connect("localhost", "user", "passwd");

if (!$conn) {

  die("Could not connect: ". mysql_error());

}

if (!mysql_select_db("database", $conn)) {

  die("Could select db: ". mysql_error());

}



// YOUR DOMAIN API KEY

$api_key = "ABCDEFGHIJK";



$query = "select * from buildings where building_latitude is null and building_longitude is null order by building_id";

$result = mysql_query($query);



while ($row = mysql_fetch_array($result)) {



  // SET ADDRESS

  $address = urlencode($row["building_street"]." ".$row["building_street_nr"]." ".$row["building_city"]." Czech republic");



  // URL TO HTTP REQUEST

  $link = "http://maps.google.com/maps/geo?q=".$address."&key=".$api_key."&sensor=false&output=csv&oe=utf8";



  // WE GET FILE CONTENT

  $page = file_get_contents($link);



  // WE OBTAIN DATA FROM GIVEN CSV

  list($status, $accuracy, $latitude, $longitude) = explode(",", $page);



  // IF EVERYTHING OK AND ACCURANCY GREATER THEN 3 WE SAVE COORDINATES

  if (($status == 200) and ($accuracy>=4)) {

    $query_edit = "update buildings set building_latitude = '".$latitude."',

    building_longitude = '".$longitude."'

    where building_id = '".$row["building_id"]."'";

    $result_edit = mysql_query($query_edit);

    echo $row["building_id"]." - OK<br />";

  } else {

    echo $row["building_id"]." - ERROR<br />";

  }



  // TIMER BECAUSE GOOGLE DOESN'T LIKE TO BE QUERIED IN SHORT TIME

  sleep(3);

}



mysql_close($conn);

?>

In following article will be introduced MySQL procedure which returns list of the nearest buildings calculated from GPS.