1,创建一个user表，包含两列name,phone

2,用python（你喜欢的任何语言）插入100W条记录（lz的笔记本比较老，大概用了1分钟吧）：

#!/usr/bin/env python
# -*- coding:utf-8 -*-

import MySQLdb

conn =  MySQLdb.connect(host='localhost',user='root',db='millionMessage')
cur = conn.cursor()

for i in range(1,1000000):
    uname = "user" + str(i)
    uphone = "188000" + str(i)
    sql = "insert into user(name,phone) values('%s','%s')" % (uname,uphone)
    cur.execute(sql)

conn.commit()
cur.close()
conn.close()

3,在没建立索引的情况下搜索：

mysql> select * from user where name='user55555';
+-------+-----------+-------------+
| uid | name | phone |
+-------+-----------+-------------+
| 55567 | user55555 | 18800055555 |
+-------+-----------+-------------+
1 row in set (0.53 sec)

mysql> select phone from user where name='user55555';
+-------------+
| phone |
+-------------+
| 18800055555 |
+-------------+
1 row in set (0.46 sec)

4,对name属性建立索引：

mysql> alter table user add index index_username(name);
Query OK, 0 rows affected (22.27 sec)
Records: 0 Duplicates: 0 Warnings: 0

5, 查询：

mysql> select * from user where name='user55555';
+-------+-----------+-------------+
| uid | name | phone |
+-------+-----------+-------------+
| 55567 | user55555 | 18800055555 |
+-------+-----------+-------------+
1 row in set (0.00 sec)

mysql> select * from user where name='user999999';
+---------+------------+--------------+
| uid | name | phone |
+---------+------------+--------------+
| 1000011 | user999999 | 188000999999 |
+---------+------------+--------------+
1 row in set (0.00 sec)

结果秒出。可见在海量数据的数据库上，索引对搜索性能的提升是非常大的。