对比mysql一般引擎与infobright--2

阅读更多

切换到同一台物理机测试
1.物理机的情况：
  内存：16g
  cpu:4核 cpu MHz : 800.000
2.建同样的库表
3.造数据5000w测试
 

	sql		数量		mysql		infobright
	在服务器上全表count		5000w		25.08s		0.08s
	python执行:d_o_lastordertime>='2012-06-19' and s_usr_level>60 and i_u_verified=1		0		36s		<1s
	count:i_u_sex=1 and s_o_usermob_type=2 and i_o_ordesnum_3m=4  and i_u_verified=1		1254437		29s;29s;29s		4s;4s;4s
	fetch_data(分页5w):i_u_sex=1 and s_o_usermob_type=2 and i_o_ordesnum_3m=4  and i_u_verified=1		50w		77s;76;75s		42s;37s;41s

4.删除重造数据10000w
   

	sql		数量		mysql		infobright
	在服务器上全表count		10000w		3m34.08s		0.06s
	python执行:d_o_lastordertime>='2012-06-19' and s_usr_level>60 and i_u_verified=1		0		36s		<1s
	count:i_u_sex=1 and s_o_usermob_type=2 and i_o_ordesnum_3m=4  and i_u_verified=1		2506968		2m55s;2m40s;2m30s		9s;9s;8s
	fetch_data(分页5w):i_u_sex=1 and s_o_usermob_type=2 and i_o_ordesnum_3m=4  and i_u_verified=1		50w		1m40s;1m41s;1m42s		1m22s;1m21s;1m22s
	fetch_data(分页5w):i_u_sex=1 and s_o_usermob_type=2 and i_o_ordesnum_3m=4  and i_u_verified=1		100w		7m22s;5m22s;1m42s		2m49s;2m49s;2m47s
	count:i_u_sex=1 and s_o_usermob_type=2 and i_o_ordesnum_3m=4  and i_u_verified=1 and f_o_total_spend_3m<200 and f_o_kdj_3m>300 and i_o_sendNumber_3m<10		49128		205s;371s;258s		29s;70s;65s
	fetch_data(分页5w):i_u_sex=1 and s_o_usermob_type=2 and i_o_ordesnum_3m=4  and i_u_verified=1		49128		285s;222s;187s		451s;449s;485s
	fetch_data(分页4个线程5w):i_u_sex=1 and s_o_usermob_type=2 and i_o_ordesnum_3m=4  and i_u_verified=1		50w		47s;40s;42s		25s;27s;25s
	fetch_data(分页4个线程5w):i_u_sex=1 and s_o_usermob_type=2 and i_o_ordesnum_3m=4  and i_u_verified=1		100w		108s;106s;107s		44s;44s;49s
	count_like(前置like查询):s_u_unicknme like '%test1'		1		330s;340s;330s		45s;44s;45s
	count_like(后置like查询):s_u_unicknme like 'test%'		10000w		108s;106s;107s		0.16s;0.01s;0.01s
	count_like(中文后置like查询):s_u_unicknme like '中午test测试%'		10000w		174s;170s;172s		0.16s;0.01s;0.01s

注：mysql分页取数据，每次的时候是递增的，infobright的时间每次相差1s摆动,数据刚加载完马上查取时，时间稍微长一点，过后测试没发现问题
1.设置统一编码utf8
    show   variables   like   "%char%";
    SHOW VARIABLES LIKE 'collation_%';
    SET NAMES 'utf8';
2.load另外张表的数据时，不影响其它表的查询