计算ctr

ctr.py

+import pandas as pd
+import pymysql
+
+df = pd.read_csv(r"/data2/models/2018-09-02predictTail6Unique.csv")
+a = eval(df.loc[0,"list"])
+a = list(map(lambda x:x[0],a))
+print(len(a))
+print(a[:2])
+cf = pd.read_csv(r"/data2/models/nvwa-2018-09-02predictTail6Unique.csv")
+b = eval(cf.loc[0,"list"])
+print(len(b))
+print(b[:2])
+a.extend(b)
+print("个数")
+print(len(set(a)))
+pre_list = list(set(a))
+pre_list = list(map(lambda x:"diary|"+x,pre_list))
+print(pre_list[:2])
+stat_date = "2018-09-02"
+cid_type = "diary"
+db = pymysql.connect(host='10.66.157.22', port=4000, user='root', passwd='3SYz54LS9#^9sBvC', db='jerry_test')
+sql = "select count(device_id) from data_feed_exposure2 \
+      where stat_date='{}' \
+      and cid_type='{}' \
+      and device_id in {}".format(stat_date,cid_type,pre_list)
+cursor = db.cursor()
+print("开始获取")
+cursor.execute(sql)
+print("成功获取")
+click = cursor.fetchall()[0][0]
+print(click)
+
+sql = "select count(device_id) from data_feed_exposure2 \
+      where stat_date='{}' \
+      and cid_type='{}' \
+      and device_id in {}".format(stat_date,cid_type,pre_list)
+cursor = db.cursor()
+print("开始获取")
+cursor.execute(sql)
+exp = cursor.fetchall()[0][0]
+print(exp)
+print(click/exp)