(1)理解 Hive 作为数据仓库在 Hadoop 体系结构中的角色。
(2)熟练使用常用的 HiveQL。
二、实验平台
- *** 作系统:Ubuntu18.04(或Ubuntu16.04);
- Hadoop版本:3.1.3;
- Hive版本:3.1.2;
- JDK版本:1.8。
三、数据集
准备工作:
-
由《Hive编程指南》(O’Reilly系列,人民邮电出版社)提供,下载地址:
https://raw.githubusercontent.com/oreillymedia/programming_hive/master/prog-hive-1st-ed-data.zip -
备用下载地址:
https://www.cocobolo.top/FileServer/prog-hive-1st-ed-data.zip -
下载慢可参考我上传的资源:林子雨Hive数据集下载
解压后可以得到本实验所需的 stocks.csv 和 dividends.csv 两个文件。
进入你的 Downloads(下载)文件夹,右键解压刚下载的数据压缩包,进入 prog-hive-1st-ed-data 文件夹,右键打开终端:
cd ~/Downloads/prog-hive-1st-ed-data sudo cp ./data/stocks/stocks.csv /usr/local/hive sudo cp ./data/dividends/dividends.csv /usr/local/hive
进入 Hadoop 目录,启动 Hadoop:
cd /usr/local/hadoop sbin/start-dfs.sh
启动 MySQL:
service mysql start
切换到 Hive 目录下,启动 MySQL 和 Hive:
cd /usr/local/hive bin/hive
四、实验步骤
(1)创建一个内部表 stocks,字段分隔符为英文逗号,表结构如下所示:
stocks 表结构:
代码:
create table if not exists stocks ( `exchange` string, `symbol` string, `ymd` string, `price_open` float, `price_high` float, `price_low` float, `price_close` float, `volume` int, `price_adj_close` float ) row format delimited fields terminated by ',';
查看表:
hive> describe stocks; OK exchange string symbol string ymd string price_open float price_high float price_low float price_close float volume int price_adj_close float Time taken: 0.062 seconds, Fetched: 9 row(s) hive>
(2)创建一个外部分区表 dividends(分区字段为 exchange 和 symbol),字段分隔符为英文逗号,表结构如下所示:
dividends 表结构
代码:
create external table if not exists dividends ( `ymd` string, `dividend` float ) partitioned by(`exchange` string ,`symbol` string) row format delimited fields terminated by ',';
查看表:
hive> describe dividends; OK ymd string dividend float exchange string symbol string # Partition Information # col_name data_type comment exchange string symbol string Time taken: 0.106 seconds, Fetched: 9 row(s) hive>
(3)从 stocks.csv 文件向 stocks 表中导入数据:
代码:
load data local inpath '/usr/local/hive/stocks.csv' overwrite into table stocks;
(4) 创建一个未分区的外部表 dividends_unpartitioned,并从 dividends.csv 向其中导入数据,表结构如下所示:
dividends_unpartitioned 表结构
代码:
create external table if not exists dividends_unpartitioned ( `exchange` string , `symbol` string, `ymd` string, `dividend` float ) row format delimited fields terminated by ',';
导入数据:
load data local inpath '/usr/local/hive/dividends.csv' overwrite into table dividends_unpartitioned;
(5)通过对 dividends_unpartitioned 的查询语句,利用 Hive 自动分区特性向分区表 dividends 各个分区中插入对应数据。
代码:
set hive.exec.dynamic.partition=true; set hive.exec.dynamic.partition.mode=nonstrict; set hive.exec.max.dynamic.partitions.pernode=1000; insert overwrite table dividends partition(`exchange`,`symbol`) select `ymd`,`dividend`,`exchange`,`symbol` from dividends_unpartitioned;
(6)查询IBM公司(symbol = IBM)从 2000 年起所有支付股息的交易日(dividends 表中有对应记录)的收盘价(price_close)。
*** 作语句如下:
select s.ymd,s.symbol,s.price_close from stocks s LEFT SEMI JOIN dividends d ON s.ymd=d.ymd and s.symbol=d.symbol where s.symbol='IBM' and year(ymd)>=2000;
输出如下(折叠部分输出):
2010-02-08 IBM 121.88 2009-11-06 IBM 123.49 2009-08-06 IBM 117.38 ... 2000-05-08 IBM 109.75 2000-02-08 IBM 118.81 Time taken: 8.75 seconds, Fetched: 41 row(s)
(7)查询苹果公司(symbol = AAPL)2008 年 10 月每个交易日的涨跌情况,涨显示 rise,跌显示 fall,不变显示 unchange。
*** 作语句如下:
select ymd, case when price_close-price_open>0 then 'rise' when price_close-price_open<0 then 'fall' else 'unchanged' end as situation from stocks where symbol='AAPL' and substring(ymd,0,7)='2008-10';
输出如下(折叠部分输出):
2008-10-31 rise 2008-10-30 rise ... 2008-10-02 fall 2008-10-01 fall Time taken: 0.1 seconds, Fetched: 23 row(s)
(8)查询 stocks 表中收盘价(price_close)比开盘价(price_open)高得最多的那条记录的交易所(exchange)、股票代码(symbol)、日期(ymd)、收盘价、开盘价及二者差价。
*** 作语句如下:
select `exchange`,`symbol`,`ymd`,price_close,price_open,price_close-price_open as `diff` from ( select * from stocks order by price_close-price_open desc limit 1 )t;
输出如下:
NASDAQ INFY 2000-02-11 670.06 534.5 135.56 Time taken: 4.476 seconds, Fetched: 1 row(s)
9)从 stocks 表中查询苹果公司(symbol=AAPL)年平均调整后收盘价(price_adj_close)大于 50 美元的年份及年平均调整后收盘价。
*** 作语句如下:
select year(ymd) as `year`, avg(price_adj_close) as avg_price from stocks where `exchange`='NASDAQ' and symbol='AAPL' group by year(ymd) having avg_price > 50;
输出如下:
2006 70.81063753105255 2007 128.27390423049016 2008 141.9790115054888 2009 146.81412711976066 2010 204.72159912109376 Time taken: 2.347 seconds, Fetched: 5 row(s)
(10)查询每年年平均调整后收盘价(price_adj_close)前三名的公司的股票代码及年平均调整后收盘价。
*** 作语句如下:
select t2.`year`,symbol,t2.avg_price from ( select *,row_number() over(partition by t1.`year` order by t1.avg_price desc) as `rank` from ( select year(ymd) as `year`, symbol, avg(price_adj_close) as avg_price from stocks group by year(ymd),symbol )t1 )t2 where t2.`rank`<=3;
输出如下(折叠部分输出):
NULL stock_symbol NULL 1962 IBM 2.0072222134423634 1962 GE 0.16876984293025638 ... 2009 GTC 174.11607115609306 2010 ISRG 319.75360107421875 2010 AMEN 313.875 2010 GTC 214.36719848632814 Time taken: 7.715 seconds, Fetched: 140 row(s)
五、总结
欢迎分享,转载请注明来源:内存溢出
评论列表(0条)