|
Geohash 是一种地址编码，它能把二维的经纬度编码成一维的字符串。比如，北海公园的编码是wx4g0ec1。
Geohash 的原理、算法
下面以(39.92324,116.3906)为例，介绍一下geohash的编码算法。
首先将纬度范围(-90,90)平分成两个区间(-90,0)、(0,90)， 如果目标纬度位于前一个区间，则编码为0，否则编码为1。由于39.92324属于(0,90)，所以取编码为1。然后再将(0,90)分成 (0,45),(45,90)两个区间，而39.92324位于(0,45)，所以编码为0。以此类推，直到精度符合要求为止，得到纬度编码为1011 1000 1100 0111 1001。
| 纬度范围 |
划分区间0 |
划分区间1 |
39.92324所属区间 |
| (-90,90) |
(-90,0.0) |
(0.0,90) |
1 |
| (0.0,90) |
(0.0,45.0) |
(45.0,90) |
0 |
| (0.0,45.0) |
(0.0,22.5) |
(22.5,45.0) |
1 |
| (22.5,45.0) |
(22.5,33.75) |
(33.75,45.0) |
1 |
| (33.75,45.0) |
(33.75,39.375) |
(39.375,45.0) |
1 |
| (39.375,45.0) |
(39.375,42.1875) |
(42.1875,45.0) |
0 |
| (39.375,42.1875) |
(39.375,40.7812) |
(40.7812,42.1875) |
0 |
| (39.375,40.7812) |
(39.375,40.0781) |
(40.0781,40.7812) |
0 |
| (39.375,40.0781) |
(39.375,39.7265) |
(39.7265,40.0781) |
1 |
| (39.7265,40.0781) |
(39.7265,39.9023) |
(39.9023,40.0781) |
1 |
| (39.9023,40.0781) |
(39.9023,39.9902) |
(39.9902,40.0781) |
0 |
| (39.9023,39.9902) |
(39.9023,39.9462) |
(39.9462,39.9902) |
0 |
| (39.9023,39.9462) |
(39.9023,39.9243) |
(39.9243,39.9462) |
0 |
| (39.9023,39.9243) |
(39.9023,39.9133) |
(39.9133,39.9243) |
1 |
| (39.9133,39.9243) |
(39.9133,39.9188) |
(39.9188,39.9243) |
1 |
| (39.9188,39.9243) |
(39.9188,39.9215) |
(39.9215,39.9243) |
1 |
经度也用同样的算法，对(-180,180)依次细分，得到116.3906的编码为1101 0010 1100 0100 0100。
| 经度范围 |
划分区间0 |
划分区间1 |
116.3906所属区间 |
| (-180,180) |
(-180,180) |
1 |
| (0.0,180) |
(0.0,90.0) |
(90.0,180) |
1 |
| (90.0,180) |
(90.0,135.0) |
(135.0,180) |
0 |
| (90.0,135.0) |
(90.0,112.5) |
(112.5,135.0) |
1 |
| (112.5,135.0) |
(112.5,123.75) |
(123.75,135.0) |
0 |
| (112.5,123.75) |
(112.5,118.125) |
(118.125,123.75) |
0 |
| (112.5,118.125) |
(112.5,115.312) |
(115.312,118.125) |
1 |
| (115.312,118.125) |
(115.312,116.718) |
(116.718,118.125) |
0 |
| (115.312,116.718) |
(115.312,116.015) |
(116.015,116.718) |
1 |
| (116.015,116.718) |
(116.015,116.367) |
(116.367,116.718) |
1 |
| (116.367,116.718) |
(116.367,116.542) |
(116.542,116.718) |
0 |
| (116.367,116.542) |
(116.367,116.455) |
(116.455,116.542) |
0 |
| (116.367,116.455) |
(116.367,116.411) |
(116.411,116.455) |
0 |
| (116.367,116.411) |
(116.367,116.389) |
(116.389,116.411) |
1 |
| (116.389,116.411) |
(116.389,116.400) |
(116.400,116.411) |
0 |
| (116.389,116.400) |
(116.389,116.394) |
(116.394,116.400) |
0 |
接下来将经度和纬度的编码合并，奇数位是纬度，偶数位是经度，得到编码 11100 11101 00100 01111 00000 01101 01011 00001。
最后，用0-9、b-z（去掉a,i,l,o）这32个字母进行base32编码，得到(39.92324,116.3906)的编码为wx4g0ec1。
| 十进制 |
0 |
1 |
2 |
3 |
4 |
5 |
6 |
7 |
8 |
9 |
10 |
11 |
12 |
13 |
14 |
15 |
| base32 |
0 |
1 |
2 |
3 |
4 |
5 |
6 |
7 |
8 |
9 |
b |
c |
d |
e |
f |
g |
| 十进制 |
16 |
17 |
18 |
19 |
20 |
21 |
22 |
23 |
24 |
25 |
26 |
27 |
28 |
29 |
30 |
31 |
| base32 |
h |
j |
k |
m |
n |
p |
q |
r |
s |
t |
u |
v |
w |
x |
y |
z |
解码算法与编码算法相反，先进行base32解码，然后分离出经纬度，最后根据二进制编码对经纬度范围进行细分即可，这里不再赘述。 不过由于geohash表示的是区间，编码越长越精确，但不可能解码出完全一致的地址。
Geohash的应用：附近地址搜索
geohash的最大用途就是附近地址搜索了。不过，从geohash的编码算法中可以看出它的一个缺点：位于格子边界两侧的两点， 虽然十分接近，但编码会完全不同。实际应用中，可以同时搜索当前格子周围的8个格子，即可解决这个问题。
最后，我们来看看本文开头提出的两个问题：速度慢，缓存命中率低。使用geohash查询附近地点，用的是字符串前缀匹配：
复制代码 代码如下: SELECT * FROM place WHERE geohash LIKE 'wx4g0%';
而前缀匹配可以利用geohash列上的索引，因此查询速度不会太慢。另外，即使用户坐标发生微小的变化， 也能编码成相同的geohash，这就保证了每次执行相同的SQL语句，使得缓存命中率大大提高。 （编辑：李大同）
【声明】本站内容均来自网络，其相关言论仅代表作者个人观点，不代表本站立场。若无意侵犯到您的权利，请及时与联系站长删除相关内容!
|
