lateral view explode使用

lateral view 側視圖   explode 爆炸               

 

select wdmc,customer_id from(
select 1 as wdmc,'14590170|14587203|21396385|22105752' as qyid
)a lateral view explode(split(qyid,'|')) as customer_id

報錯：

Error while compiling statement: FAILED: ParseException line 5:44 extraneous input 'customer_id' expecting EOF near '<EOF>'

側視圖本質是left join，  explode 之後就是 右表了，原表爲座標

而左右兩表 都需要  命名

正確的寫法： 多了一個b（側視圖命名） 而已

select wdmc,customer_id from(
select 1 as wdmc,"14590170|14587203|21396385|22105752" as qyid
)a lateral view explode(split(qyid,"|"))b as customer_id

但結果有問題，我想用‘|’隔開  但結果是：

	wdmc	customer_id
1	1	1
2	1	4
3	1	5
4	1	9
5	1	0
6	1	1
7	1	7
8	1	0
9	1	|
10	1	1
11	1	4
12	1	5
13	1	8

...

嘗試單獨使用explode：

select explode(split("14590170|14587203|21396385|22105752","|")) as id

 結果一樣：

	id
1	1
2	4
3	5
4	9
5	0
6	1
7	7
8	0
9	|
10	1
11	4
12	5
13	8
14	7
15	2

split有問題 

describe function split

	tab_name
1	split(str, regex) - Splits str around occurances that match regex

後來慢慢分析，原來是轉義有問題，單獨一個  ‘|’   在 hive中可能是別的含義，我們需要轉義‘|’ 讓hive知道我要的就是 ‘|’用它，

但是  ‘\|’ 其實這樣也是錯的，因爲 ‘\’ 在hive中也有別的意思，所以，我們要將斜槓  ‘\’  先轉義成  真斜槓

select explode(split("14590170|14587203|21396385|22105752","\\|")) as id

 

	id
1	14590170
2	14587203
3	21396385
4	22105752

再搭配 側視圖  lateral view 

select wdmc,customer_id from(
select 1 as wdmc,"14590170|14587203|21396385|22105752" as qyid
)a lateral view explode(split(qyid,"\\|"))b as customer_id

 

結果完美：

	wdmc	customer_id
1	1	14590170
2	1	14587203
3	1	21396385
4	1	22105752