shell 脚本执行python脚本，连接hive提交数据写入表

使用说明

1.cd /opt/zy
在这个目录下以root用户权限执行命令
2.
在SAP查询的时候
Tcode:ZMMR0005
Purchase Org *
PO Creating：2017/3/1 (开始日期） 2017/6/31(结束日期）
Vendor
1000341
plant *

这样查询处理的结果代表 发货日期在20170301-20170631的所有记录，不管到达日期在那个月

从SAP导出数据表格，存为txt形式以”\t”分隔
用rz命令把导出的文件上传到/opt/zy目录下，
3.执行命令 注意参数必须严格符合XXXXXXXXtoYYYYYYYY的格式,代表startdate to enddate
example:
[root@slave1 zy]# bash try2.sh 20170301to20170632
4.去Hue里查询分析结果
SELECT * from saplifttime WHERE querypocredatestart=’XXXXXXXX’[and querypocredateend=’YYYYYYYY’];

5.如果想看原数据，去pcg.sap表，命令如下:
SELECT * from sap WHERE querypocredatestart=’20170301’;

运行结果截图：

技术实现说明

用shell 脚本调用python脚本

shell 脚本 try2.sh

#!/bin/sh
#echo $1
daterange=$1#赋值给daterange这个变量是因为后面截取字符串要用到，否则我不会写
python3 /opt/zy/runtask.py $1 #运行python脚本
startdate=${daterange:0:8}   #截取查询的开始日期
#echo $startdate
enddate=${daterange:10:18}   #截取查询的结束日期
#echo $enddate
sed -i '1,3d' /opt/zy/$1.txt   #删除前三行，因为前三行是空行
sed 's/.\{1\}//' $1.txt>$1regular.txt #删除第一列，因为第一列是空列
hdfs dfs -put -f /opt/zy/$1regular.txt /user/hive/pcg-data/zhouyi6_files #把服务器上的本地文件上传到hadoop集群上
hive -e "LOAD DATA INPATH '/user/hive/pcg-data/zhouyi6_files/$1regular.txt' INTO TABLE pcg.sap partition(querypocredatestart=$startdate,querypocredateend=$enddate)" #把文件的数据载入表 
rm $1.txt #删除本地原文件，只保留格式处理后的文件

备注：
1.因为sed命令 不修改文件本身，所以要把修改后的结果存入新文件 +regular后缀的
2.sed -i，-i代表不把删除前三行后的结果显示在命令行上
3.hdfs dfs -put -f
-f option will overwrite the destination if it already exists.
4.运行这个脚本的前提是已经创建了pcg.sap表，建表语句如下：

CREATE TABLE SAP(`PO Cre Date` string,
`Vendor` string, 
`WW Partner` string, 
`Name of Vendor` string,
`PO Cre by` string, 
`Purch Doc Type` string,
`Purch Order` string,
`PO Item` string,
`Deletion Indicator in PO Item` string, 
`Request Shipment Day` string,
`Material` string,
`Short Text` string, 
`Plant` string, 
`Issuing Stor location` string,
`Receive Stor loaction` string, 
`PO item change date` string, 
`Delivery Priority` string,
`PO Qty` string,
`Total GR Qty` string,
`Still to be delivered` string,
`Delivery Note` string,
`Delivery Note Type (ASN or DN)` string, 
`Delivery Note item` string,
`Delivery Note qty` string, 
`Delivery Note Creation Date` string,
`Delivery Note ACK Date` string, 
`Incoterm` string, 
`Part Battery Indicator` string,
`BOL/AWBill` string, 
`Purchase order type` string, 
`Gr Date`string) 
partitioned by (`queryPoCreDateStart` string,`queryPoCreDateEnd` string)
row format delimited fields terminated by "\t" stored as textfile

python脚本

import  pandas as pd
import  sys
data = pd.read_csv(sys.argv[1]+".txt", sep="\t")
#print(data.columns)
data['Delivery Note Creation Date']=pd.to_datetime(data['Delivery Note Creation Date'],format='%d.%m.%Y')
data['Gr Date']=pd.to_datetime(data['Gr Date'],format='%d.%m.%Y')
data=data.drop(data[data['Delivery Note Creation Date'].isnull()].index.tolist())#删除某列为空值所在的行
data=data.drop(data[data['Gr Date'].isnull()].index.tolist())#删除某列为空值所在的行
data['delta']=(data['Gr Date']-data['Delivery Note Creation Date']).apply(lambda  x:x.days)#相差的时间
print(data['delta'].describe())
#sql_content="insert into table saplifttime values(%,%s,%s,%s,%s,%s,%s,%s,%s,%s)"%\
import hdfs
from impala.dbapi import connect
filename=sys.argv[1]+".txt"
hdfspath='/user/hive/pcg-data/zhouyi6_files'
client=hdfs.Client("http://10.100.208.222:50070")#50070
#8888是我登录WEB 操作界面时候的接口
#print(client.status("/user/zhouyi",strict=True))#查看路径信息
#print(client.list("/user/zhouyi"))#查看文件夹下的文件
#client.upload(hdfs_path=hdfspath,local_path="/opt/zy/"+filename,overwrite=True)
# overwrtie=True means Delete any uploaded files if an error occurs during theupload.
conn = connect(host='10.100.208.222', port=21050,database='pcg')
cur = conn.cursor()
stdate,edate=sys.argv[1].split("to")
#print(sys.argv[1])
cnt=str(data['delta'].describe()[0])
mean=str(data['delta'].describe()[1])
std=str(data['delta'].describe()[2])
mini=str(data['delta'].describe()[3])
twentyfive=str(data['delta'].describe()[4])
fifty=str(data['delta'].describe()[5])
seventyfive=str(data['delta'].describe()[6])
maxm=str(data['delta'].describe()[7])
args=[stdate,edate,cnt,mean,std,mini,twentyfive,fifty,seventyfive,maxm]
print(args)

#对的SQL
#sql_content="insert into table saplifttime values("+str(5555)+",'20200607','22','4.2','9.88','1','2','5','10','9999999999999')"
sql_content="insert into table saplifttime values(?,?,?,?,?,?,?,?,?,?)"
cur.execute(sql_content,args)#把运算结果插到表pcg.saplifttime里

备注：
1.执行cur.execute的前提是已经建好pcg.saplifttime的表，建表语句如下：

CREATE TABLE SAPLifttime(querypocredatestart STRING,querypocredateend STRING,cnt STRING,mean STRING,std STRING,minimum STRING,25percent STRING,50percent STRING,75percent STRING,maxmum STRING)
ROW FORMAT DELIMITED FIELDS TERMINATED BY "\t" STORED AS Textfile

2.计算逻辑：
第一步：把字段“Delivery Note Creation Date”视为货物发出日期，如果为空则删除该行
第二步：把字段“Gr Date”视为货物到达日期，如果为空则删除该行
第三步：货物在途时间= Gr Date - Delivery Note Creation Date
第四步：对货物在途时间求cnt,mean,std,minimum,25%,50%,75%,maxmum

踩过的坑：
1.我的表字段都是STRING类型，values的占位符问题，我一开始试过%s,%d,总与python里对应的值格式不匹配。后来用?占位就好了
2.cur.excute(sql.args)这样写的好处在于看起来清晰，不用拼接特别长的sql字符串了，非常容易拼错