技術標籤：OLAP--ClickHousehadoop

– with語法

 with '1' as v select * from tb_user where id=v;
 --求平均年齡
 with (select count(1) from tb_user) as cnt select (sum(age)/cnt) from tb_user;

array join 語法

array join 相當於explode+lateral view

select id,arr from tb_arr_join array join arr;  --會覆蓋原先的陣列
select id,arr,
 
x from tb_arr_join array join arr as x ;  
 select id,arr,x from tb_arr_join array join [1,2,3,4,5] as x; --自動拼
  ┌─id─┬─arr──────────────┬─x─┐
│  1 │ ['a1','a2']      │ 1 │
│  1 │ ['a1','a2']      │ 2 │
│  1 │ ['a1','a2']      │ 3 │
│  1 │ ['a1','a2']      │ 4 │
│  1 │ ['a1','a2']      │ 5 │
│  2 │ [
 
'b1','b2','b3'] │ 1 │
│  2 │ ['b1','b2','b3'] │ 2 │
│  2 │ ['b1','b2','b3'] │ 3 │
│  2 │ ['b1','b2','b3'] │ 4 │
│  2 │ ['b1','b2','b3'] │ 5 │
└────┴──────────────────┴───┘

案例需求：將如下資料成功轉換


程式碼實現：

select id ,e,i from (select id ,groupArray(name) arr, 
arrayEnumerate(arr) arr_index
from tb_test_arr
group
 
 by id ) t 
array join arr as e,
arr_index as i
;

FORMAT 指定輸出和輸入的資料格式

clickhouse-client  -q  "select * from tb_user  FORMAT  XML" 

LIMIT/LIMIT BY 語法

select * from tb_user limit 2,2 ; --從索引為第二個開始找兩個

程式碼實現：

select * from tb_limit limit 2 by name; --通過name分組

create  view  v_limit as   select * from tb_limit ;   --建立一個檢視臨時儲存資料  不會儲存到磁碟
create  table  t_limit engine=Log as   select * from tb_limit ;
create  table  tb_limit2  like  tb_limit ;  --不支援

建立分割槽表

按照指定欄位分組

create table  tb_p(
oid String ,
money  Float64 ,
cDate  Date
)  engine = MergeTree 
order by oid 
partition by cDate ;  
insert into  tb_p values ('002',99,'2020-12-01') ,('001',98,'2020-12-01') ,('003',199,'2020-12-02');

┌─oid─┬─money─┬──────cDate─┐
│ 003 │   199 │ 2020-12-02 │
└─────┴───────┴────────────┘
┌─oid─┬─money─┬──────cDate─┐
│ 001 │    98 │ 2020-12-01 │
│ 002 │    99 │ 2020-12-01 │
----------------------------

按照月份分組

create table  tb_p2(
oid String ,
money  Float64 ,
cDate  Date
)  engine = MergeTree 
order by oid 
partition by toMonth(cDate) ;
insert into  tb_p2 values ('002',99,'2020-12-01') ,('001',98,'2020-12-01') ,('003',199,'2020-12-02'),('004',299,'2020-11-02');

┌─oid─┬─money─┬──────cDate─┐
│ 001 │    98 │ 2020-12-01 │
│ 002 │    99 │ 2020-12-01 │
│ 003 │   199 │ 2020-12-02 │
└─────┴───────┴────────────┘
┌─oid─┬─money─┬──────cDate─┐
│ 004 │   299 │ 2020-11-02 │
└─────┴───────┴────────────┘

指定兩個分割槽欄位 年+月 不同就分割槽！！！！

create table  tb_p3(
oid String ,
money  Float64 ,
cDate  Date
)  engine = MergeTree 
order by oid 
partition by (toYear(cDate) , toMonth(cDate)) ;-- 月進行分割槽 
insert into  tb_p3 values ('002',99,'2020-12-01') ,('001',98,'2020-12-01') ,('003',199,'2020-12-02'),('004',299,'2020-11-02'),('005',299,'2019-11-02')
 ;
drwxr-x---. 2 clickhouse clickhouse 228 Dec  6 04:33 2019-11_3_3_0
drwxr-x---. 2 clickhouse clickhouse 228 Dec  6 04:33 2020-11_2_2_0
drwxr-x---. 2 clickhouse clickhouse 228 Dec  6 04:33 2020-12_1_1_0