fix

azkaban/app_start.sh

@@ -2,6 +2,6 @@
 
 DT=$(date +%Y%m%d -d "-1 day $ScheduleTime")
 
-LAST_DT=$(date +%Y%m%d -d "-7 day $DT")
+LAST_DT=$(date +%Y%m%d -d "-30 day $DT")
 
 hive -hivevar dt=${DT} -hivevar last_dt=${LAST_DT} -f app_start.sql
\ No newline at end of file

azkaban/app_start.sql

@@ -5,10 +5,13 @@ SET mapreduce.map.memory.mb=4096;
 SET mapreduce.map.java.opts=-Xmx8192m;
 SET mapreduce.reduce.memory.mb=4096;
 SET mapreduce.reduce.java.opts=-Xmx8192m;
+SET hive.mapred.mode=nonstrict;
 
 DROP TABLE IF EXISTS package_num;
-CREATE TABLE package_num AS SELECT package_name,count(1) install_num FROM dwh.dm_install_list_v2 WHERE dt = '${dt}' GROUP BY package_name
-    HAVING count(1) > 30000 ORDER BY install_num DESC LIMIT 5000;
+CREATE TABLE package_num AS SELECT d.package_name,count(1) install_num FROM dwh.dm_install_list_v2 d,dwh.dim_package_tags t
+    WHERE d.dt = '${dt}' AND d.update_date BETWEEN '${LAST_DT}' AND '${dt}' AND d.package_name <> t.package_name
+    GROUP BY d.package_name HAVING count(1) > 30000
+    ORDER BY install_num DESC LIMIT 5000;
 
 DROP TABLE IF EXISTS package_info;
 CREATE TABLE package_info AS