explode(官网链接)
explode 是一个 UDTF(表生成函数),将单个输入行转换为多个输出行。一般和 lateral view 结合使用,主要有两种用法:
输入类型 | 使用方法 | 描述 |
T | explode(ARRAY<T> a) | 将数组分解为多行,返回单列多行,每一行代表数组的一个元素 |
Tkey,Tvalue | explode(MAP<Tkey,Tvalue> m) | 将 MAP 分解为多行,返回的行具有两列(键-值),每一行代表输入中的一个键值对 |
示例:
explode(array)
hive (default)> select explode(array('A','B','C'));
OK
col
A
B
C
Time taken: 0.402 seconds, Fetched: 3 row(s)
hive (default)> select explode(array('A','B','C')) as col1;
OK
col1
A
B
C
Time taken: 0.145 seconds, Fetched: 3 row(s)
hive (default)> select tf.* from (select 0) t lateral view explode(array('A','B','C')) tf;
OK
tf.col
A
B
C
Time taken: 0.191 seconds, Fetched: 3 row(s)
hive (default)> select tf.* from (select 0) t lateral view explode(array('A','B','C')) tf as col1;
OK
tf.col1
A
B
C
explode(map)
hive (default)> select explode(map('A',10,'B',20,'C',30));
OK
key value
A 10
B 20
C 30
Time taken: 0.153 seconds, Fetched: 3 row(s)
hive (default)> select explode(map('A',10,'B',20,'C',30)) as (my_key,my_value);
OK
my_key my_value
A 10
B 20
C 30
Time taken: 0.137 seconds, Fetched: 3 row(s)
hive (default)> select tf.* from (select 0) t lateral view explode(map('A',10,'B',20,'C',30)) tf;
OK
tf.key tf.value
A 10
B 20
C 30
Time taken: 0.128 seconds, Fetched: 3 row(s)
hive (default)> select tf.* from (select 0) t lateral view explode(map('A',10,'B',20,'C',30)) tf as my_key,my_value;
OK
tf.my_key tf.my_value
A 10
B 20
C 30
Time taken: 0.109 seconds, Fetched: 3 row(s)
LateralView(官网链接)
语法
lateralView: LATERAL VIEW udtf(expression) tableAlias AS columnAlias (',' columnAlias)*
fromClause: FROM baseTable (lateralView)*
描述
lateralview 与用户自定义表生成函数(UDTF)(例如 explode)结合使用,UDTF 为每个输入行生成零个或多个输出行。lateralview 首先将 UDTF 应用于基础表的每一行,然后将结果输出行与输入行连接起来形成具有所提供表别名的虚拟表。
实例
基础表 pageads 具有两列:pageid(页面名称)和 add_list(页面上显示的广告数组)。
hive (test)> !cat pageads
> ;
front_page 1,2,3
contact_page 3,4,5
hive (test)> create table pageads(
> pageid string,
> add_list array<int>)
> ROW FORMAT delimited
> fields terminated by '\t'
> collection items terminated by ','
> lines terminated by '\n'
> ;
OK
Time taken: 0.099 seconds
hive (test)> load data local inpath '/home/hadoop/pageads' into table pageads;
Loading data to table test.pageads
OK
Time taken: 0.331 seconds
hive (test)> select * from pageads;
OK
pageads.pageid pageads.add_list
front_page [1,2,3]
contact_page [3,4,5]
Time taken: 0.106 seconds, Fetched: 2 row(s)
hive (test)> select pageid,addid from pageads lateral view explode(add_list) adTable as addid;
OK
pageid addid
front_page 1
front_page 2
front_page 3
contact_page 3
contact_page 4
contact_page 5
Time taken: 0.105 seconds, Fetched: 6 row(s)
hive (test)> select addid,count(1) from pageads lateral view explode(add_list) adTable as addid group by addid;
.......
OK
addid _c1
1 1
2 1
3 2
4 1
5 1
多个lateralview
from 子句可以具有多个 lateralview 子句。后续的 lateralview 子句可以引用 lateralview 左侧表中的任何列。例如以下查询:
SELECT * FROM exampleTable
LATERAL VIEW explode(col1) myTable1 AS myCol1
LATERAL VIEW explode(myCol1) myTable2 AS myCol2;
注意,lateralview 子句按其出现的顺序应用。
实例
hive (test)> !cat basetable;
1,2 a,b,c
3,4 d,e,f
hive (test)> create table basetable(
> col1 array<int>,
> col2 array<string>)
> ROW FORMAT delimited
> fields terminated by '\t'
> collection items terminated by ','
> lines terminated by '\n'
> ;
OK
Time taken: 0.113 seconds
hive (test)> load data local inpath '/home/hadoop/basetable' into table basetable;
Loading data to table test.basetable
OK
Time taken: 0.329 seconds
hive (test)> select * from basetable;
OK
basetable.col1 basetable.col2
[1,2] ["a","b","c"]
[3,4] ["d","e","f"]
Time taken: 0.104 seconds, Fetched: 2 row(s)
hive (test)> SELECT myCol1, col2 FROM basetable
> LATERAL VIEW explode(col1) myTable1 AS myCol1;
OK
mycol1 col2
1 ["a","b","c"]
2 ["a","b","c"]
3 ["d","e","f"]
4 ["d","e","f"]
Time taken: 0.089 seconds, Fetched: 4 row(s)
hive (test)> SELECT myCol1, myCol2 FROM baseTable
> LATERAL VIEW explode(col1) myTable1 AS myCol1
> LATERAL VIEW explode(col2) myTable2 AS myCol2;
OK
mycol1 mycol2
1 a
1 b
1 c
2 a
2 b
2 c
3 d
3 e
3 f
4 d
4 e
4 f
Time taken: 0.093 seconds, Fetched: 12 row(s)
Outer Lateral Views
hive (test)> SELECT * FROM basetable LATERAL VIEW explode(array()) C AS a limit 10;
OK
basetable.col1 basetable.col2 c.a
Time taken: 0.063 seconds
hive (test)> SELECT * FROM basetable LATERAL VIEW OUTER explode(array()) C AS a limit 10;
OK
basetable.col1 basetable.col2 c.a
[1,2] ["a","b","c"] NULL
[3,4] ["d","e","f"] NULL
Time taken: 0.092 seconds, Fetched: 2 row(s)