temporary table /view 探究

explain
select
    *
from t1

union all
select
    *
from t1

view 每次执行都会有新的值产生

 

从 执行计划上看view 就是用的不同的值参与随机

== Physical Plan ==
Union
:- *(1) Project [random(-3323905141429424159) AS a#44]
:  +- *(1) Scan OneRowRelation[]
+- *(2) Project [random(5605563624691597816) AS a#46]
   +- *(2) Scan OneRowRelation[]

------------------------------------------------  temporary table

create temporary  table  t1 as (
    select random() as a
);
-- explain
select
    *
from t1

union all
select
    *
from t1

查看执行计划拿到的是一个临时的数据, 固定的结果值

== Physical Plan ==
Union
:- Scan hive work_table.spark_temp_application_1677888387085_9802_t1 [a#80], HiveTableRelation [`work_table`.`spark_temp_application_1677888387085_9802_t1`, org.apache.hadoop.hive.serde2.lazy.LazySimpleSerDe, Data Cols: [a#80], Partition Cols: []]
+- Scan hive work_table.spark_temp_application_1677888387085_9802_t1 [a#81], HiveTableRelation [`work_table`.`spark_temp_application_1677888387085_9802_t1`, org.apache.hadoop.hive.serde2.lazy.LazySimpleSerDe, Data Cols: [a#81], Partition Cols: []]

总结 临时表会落表数据, view 重用的是代码段.