redis作为内存数据库,常被用于应对实时响应的高并发缓存。
之前的推荐引擎程序将推荐结果写入mysql,全部写完后再发更新redis请求。一个java程序接收http响应然后开始更新redis,受限于mysql的IO,这一步速度极慢。准备改用python将推荐结果直接写入redis,然后同步写mysql做持久化保存。
推荐结果是一个列表,采用的是redis的序列操作。
import redis
import simplejson as sj
r=redis.Redis(host=redishost,port=6379,db=0)
# print r.dbsize()
# print sj.dumps({"assetId":"","assetname":"不二神探","assetpic":"","movieID":"180353","providerId":"","rank":"1"})
# print 'list index 0:',r.lindex('Topcontentlist_3_50047',0)
# print 'list lrange:',r.lrange('Topcontentlist_3_50047',start=0,end=-1)
strs = sj.dumps({"assetId":"","assetname":"不二神探","assetpic":"","movieID":"180353","providerId":"","rank":"1"})
strb = sj.dumps({"assetId":"","assetname":"二神探","assetpic":"","movieID":"1803","providerId":"","rank":"2"})
r.lpush('s',strs)
r.lpush('s',strb)
print 'list index 0:',r.lindex('s',0)
print 'list lrange:',r.lrange('s',start=0,end=-1)
print r.lpop('s')
print 'list lrange:',r.lrange('s',start=0,end=-1)
r.delete('s')
r.lpush将strs这个字符串推到's'这个key下;
r.lindex('s',0)是在's'这个key下查询第一个值;同理,r.lindex('s',1)是第二个值,没有值返回None;
r.lpop('s')是pop出第一个项;
r.lrange()批量查询;
r.delete()删除该Key-Value
list index 0: {"movieID": "1803", "assetId": "", "rank": "2", "assetpic": "", "providerId": "", "assetname": "\u4e8c\u795e\u63a2"}
list lrange: ['{"movieID": "1803", "assetId": "", "rank": "2", "assetpic": "", "providerId": "", "assetname": "\\u4e8c\\u795e\\u63a2"}', '{"movieID": "180353", "assetId": "", "rank": "1", "assetpic": "", "providerId": "", "assetname": "\\u4e0d\\u4e8c\\u795e\\u63a2"}']
{"movieID": "1803", "assetId": "", "rank": "2", "assetpic": "", "providerId": "", "assetname": "\u4e8c\u795e\u63a2"}
list lrange: ['{"movieID": "180353", "assetId": "", "rank": "1", "assetpic": "", "providerId": "", "assetname": "\\u4e0d\\u4e8c\\u795e\\u63a2"}']