需求描述:
一般的统计对格式如:”dim1,dim2,index1,index2,date“,如果维度对应的指标不存在,则结果表中就没有这个记录。现在要求,指标为空的记录给补上,且将其指标置为0。
实现代码:
/Users/nisj/PycharmProjects/EsDataProc/result_null_proc.py
# -*- coding=utf-8 -*-
import os
import re
nrpd_list = []
for nrp_list in result_data:
nrp = re.split('\t', nrp_list.replace('\n', ''))
nrpd_list.append(nrp)
for nrpd in nrpd_list:
appkey = nrpd[0]
appsource = nrpd[1]
week_count = nrpd[2]
result_data1 = os.popen("""mysql -hMysqlHost -P6603 -uhadoop -pMysqlPass -N -e "use funnyai_data; \
select pm,appkey,app_source,week_count,compare_week_count from \
(SELECT (SELECT COUNT(1) + 1 \
FROM (select distinct remain_week from bi_user_remain_pay_byweek ) A \
WHERE concat(substr(A.remain_week,1,4),lpad(substr(A.remain_week,6),2,0)) < concat(substr(B.remain_week,1,4),lpad(substr(B.remain_week,6),2,0))) pm \
FROM (select distinct remain_week from bi_user_remain_pay_byweek ) B \
ORDER BY PM) a1 \
left join ( \
select appkey,app_source,week_count,compare_week_count \
from x_chushou_user_compare_week_stat \
where appkey='%s' and app_source='%s' and week_count='%s' \
) a2 on a1.pm=a2.compare_week_count \
where a1.pm>'%s' and a2.appkey is null \
order by a1.pm;" """ % (appkey, appsource, week_count, week_count)).readlines();
nrpd_list = []
for nrp_list in result_data1:
nrp = re.split('\t', nrp_list.replace('\n', ''))
nrpd_list.append(nrp)
for nrpd in nrpd_list:
pm = nrpd[0]
# print pm,appkey,appsource,week_count
os.system("""mysql -hMysqlHost -P6603 -uhadoop -pMysqlPass -N -e "use funnyai_data; \
insert into x_chushou_user_compare_week_stat(appkey,app_source,week_count,compare_week_count,compare_identify_count,compare_register_user_count,compare_user_count,compare_pay_amount,compare_pay_user_count) \
select '%s','%s','%s','%s','%d','%d','%d','%d','%d'; \
" """ % (appkey, appsource, week_count, pm, 0, 0, 0, 0, 0))
resultNullProc()