最近在用Hive做多维数据分析，总结一些常用HiveQL命令。
1. 建表
以纯文本数据建表：
create table default.calendar_table
(
day_cal date
,week_cal string
,montn_cal string
,year_cal string
)
row format delimited fields terminated by ','
stored as textfile;
若未指定为外部表（external table），则默认为托管表（managed table）。二者的区别在于load与drop操作：托管表用load data inpath加载数据（路径可为本地目录，也可是HDFS目录），该操作会将该文件放在HDFS目录：/user/hive/warehouse/ 下；而外部表的数据是在location中指定，一般配合partition描述数据的生成信息；drop托管表时会将元数据与/user/hive/warehouse/下的数据一起删掉，而drop外部表时只会删除元数据。将本地文件加载到托管表：
load data local inpath 'cal.csv' overwrite into table default.calendar_table;
以orc file数据建外部表表：
create external table default.ad_base
(
uid string
,adx string
,exposure string
,click string
)
partitioned by (day_time date)
stored as orc
location '/<hdfs>/<path>';
2. Partition
增加partition并指定location：
alter table DEFAULT.ad_base
add if not exists partition (day_time=date '2016-05-20')
location '2016-05-20/xxx';
重新设置partition的location：
alter table DEFAULT.ad_base
partition (day_time=date '2016-05-20')
set location 'hdfs://<path>/<to>/';
删除partition
alter table DEFAULT.ad_base
drop if exists partition (day_time=date '2016-05-20')
ignore protection;
查看所有的paritition，以及查看某一partition的详细信息：
show partitions ad_base;
describe formatted ad_base partition(day_time = '2016-05-20');
3. UDF
Hive的UDF非常丰富，基本能满足大部分的需求。
正则匹配获取相应字符串：
regexp_extract(b.dvc, '(.*)_(.*)', 2) as imei
复杂数据类型map、struct、指定schema的struct、array、union的构造如下：
map(key1, value1, key2, value2, ...)
struct(val1, val2, val3, ...)
named_struct(name1, val1, name2, val2, ...)
array(val1, val2, ...)
create_union(tag, val1, val2, ...)
获取复杂数据类型的某列值：
array: A[n]
map: M[key]
struct: S.x
条件判断case when，比如，在left join中指定默认值：
select uid, media,
case when b.tag is NULL then array(named_struct('tag','EMPTY', 'label','EMPTY')) else b.tag end as tags
from ad_base a
left outer join ad_tag b on (a.uid = regexp_extract(b.dvc, '(.*)_(.*)', 2) and exposure = '1');
4. UDTF
UDTF主要用来对复杂数据类型进行平铺操作，比如，explode平铺array与map，inline平铺array<struct>；这种内置的UDTF要与lateral view配合使用：
select myCol1, col2 FROM baseTable
lateral view explode(col1) myTable1 AS myCol1;
select uid, tag, label
from ad_tag
lateral view inline(tags) tag_tb;
5. 多维分析
Hive 提供grouping set、rollup、cube关键字进行多维数据分析，可以解决自定义的维度组合、上钻维度（n+1n+1种）组合、所有的维度组合（2n2n种）的需求。比如：
SELECT a, b, SUM( c )
FROM tab1
GROUP BY a, b GROUPING SETS ( (a, b), a, b, ( ) )
SELECT a, b, SUM( c ) FROM tab1 GROUP BY a, b
UNION
SELECT a, null, SUM( c ) FROM tab1 GROUP BY a, null
UNION
SELECT null, b, SUM( c ) FROM tab1 GROUP BY null, b
UNION
SELECT null, null, SUM( c ) FROM tab1
GROUP BY a, b, c, WITH ROLLUP
GROUP BY a, b, c GROUPING SETS ( (a, b, c), (a, b), (a), ( ))
GROUP BY a, b, c WITH CUBE
GROUP BY a, b, c GROUPING SETS ( (a, b, c), (a, b), (b, c), (a, c), (a), (b), (c), ( ))
此外，Hive还提供了GROUPING__ID函数对每一组合的维度进行编号，以区分该统计属于哪一维度组合，比如：
select tag, media, grouping__id, count(*) as pv
from ad_base
group by tag, media with rollup;
以指定分隔符保存结果到本地目录：
explain
INSERT OVERWRITE LOCAL DIRECTORY '/home/<path>/<to>'
ROW FORMAT DELIMITED
FIELDS TERMINATED BY '\t'
select media, count(distinct uid) as uv
from ad_base
where day_time = '2016-05-20' and exposure = '1'
group by media;
