场景:
hive有张表armmttxn_tmp,其中有一个字段lot_number,该字段以逗号分隔开多个值,每个值又以冒号来分割料号和数量,如:A3220089:-40,A3220090:-40,A3220091:-40,A3220083:-40,A3220087:-40,A3220086:-40,A3220088:-40,A3220084:-40,A3220081:-40,A3220082:-40,A3220092:-40,A3220093:-40,A3220085:-40,A3220094:-40。
要求:
把该字段拆分开来,并且把料号和数量单独列出,也就是分成两列。
原数据:
sql
select
key_id,
lot_number
from armmttxn_tmp
where key_id = '48641906';
用到的函数:split()、explode()
步骤:
step1:以逗号拆分开,如下:
["A3220089:-40","A3220090:-40","A3220091:-40","A3220083:-40","A3220087:-40","A3220086:-40","A3220088:-40","A3220084:-40","A3220081:-40","A3220082:-40","A3220092:-40","A3220093:-40","A3220085:-40","A3220094:-40"]
使用split函数,把数据拆分开
sql
select
key_id ,
split(lot_number, ',') lot_number
from armmttxn_tmp
where key_id = '48641906';
step2:一行变成多行
sql
select
explode(split(lot_number, ',')) lot_number
from armmttxn_tmp
where key_id = '48641906';
这里如果加上key_id字段,会怎样呢?
SQL 错误 [10081] [42000]: Error while compiling statement: FAILED: SemanticException [Error 10081]: UDTF's are not supported outside the SELECT clause, nor nested in expressions
原因:当使用UDTF函数的时候,hive只允许对拆分字段进行访问。
所以,可以这样使用:select explode(split(lot_number, ',')) lot_number from armmttxn_tmp where key_id = '48641906';
但不可以这样使用:select key_id ,explode(split(lot_number, ',')) lot_number from armmttxn_tmp where key_id = '48641906';
如果想访问除了拆分字段以外 的字段,怎么办呢?
用lateral view侧视图!
lateral view为侧视图,是为了配合UDTF来使用,把某一行数据拆分成多行数据.不加lateral view的UDTF只能提取单个字段拆分,并不能塞会原来数据表中.加上lateral view就可以将拆分的单个字段数据与原始表数据关联上.
注意:在使用lateral view的时候需要指定视图别名
--表名 lateral view UDTF(xxx) 视图别名(虚拟表名) as a,b,c(列别名)
--lateral view explode 相当于一个拆分lot_number字段的虚表,然后与原表进行关联.
step3:拆分的字段与原始表数据关联上.
sql
select
key_id ,
split(view.*,':') lot_number
from armmttxn_tmp lateral view explode(split(lot_number, ',')) view
where key_id = '48641906';
但还不是我们想要的最终结果,还需要把lot_number拆分成两列
step4: 拆分成两列
sql
select
key_id ,
split(view.*,':')[size(split(view.*, ':'))-2] as lot_number,
split(view.*,':')[size(split(view.*, ':'))-1] as quantity
from armmttxn_tmp lateral view explode(split(lot_number, ',')) view
where key_id = '48641906';