Hive explain执行计划详解
简介:HIVE提供了EXPLAIN命令来展示一个查询的执行计划,这个执行计划对于我们了解底层原理,hive 调优,排查数据倾斜等很有帮助
一、EXPLAIN 参数介绍
语法 :
EXPLAIN [EXTENDED|CBO|AST|DEPENDENCY|AUTHORIZATION|LOCKS|VECTORIZATION|ANALYZE] querySql
EXTENDED:加上 extended 可以输出有关计划的额外信息。这通常是物理信息,例如文件名。这些额外信息对我们用处不大
CBO:输出由Calcite优化器生成的计划。CBO 从 hive 4.0.0 版本开始支持
AST:输出查询的抽象语法树。AST 在hive 2.1.0 版本删除了,存在bug,转储AST可能会导致OOM错误,将在4.0.0版本修复
DEPENDENCY:dependency在EXPLAIN语句中使用会产生有关计划中输入的额外信息。它显示了输入的各种属性
AUTHORIZATION:显示所有的实体需要被授权执行(如果存在)的查询和授权失败
LOCKS:这对于了解系统将获得哪些锁以运行指定的查询很有用。LOCKS 从 hive 3.2.0 开始支持
VECTORIZATION:将详细信息添加到EXPLAIN输出中,以显示为什么未对Map和Reduce进行矢量化。从 Hive 2.3.0 开始支持
ANALYZE:用实际的行数注释计划。从 Hive 2.2.0 开始支持
二、简单sum例子
2.1 执行计划查询Sql和结果
explain select sum(id) from dw.ods_bdg_db_statistics_compass_property where dt='20220627';
STAGE DEPENDENCIES:
Stage-1 is a root stage
Stage-0 depends on stages: Stage-1
STAGE PLANS:
Stage: Stage-1
Map Reduce
Map Operator Tree:
TableScan
alias: ods_bdg_db_statistics_compass_property
Statistics: Num rows: 7794 Data size: 31177 Basic stats: COMPLETE Column stats: NONE
Select Operator
expressions: id (type: int)
outputColumnNames: id
Statistics: Num rows: 7794 Data size: 31177 Basic stats: COMPLETE Column stats: NONE
Group By Operator
aggregations: sum(id)
mode: hash
outputColumnNames: _col0
Statistics: Num rows: 1 Data size: 8 Basic stats: COMPLETE Column stats: NONE
Reduce Output Operator
sort order:
Statistics: Num rows: 1 Data size: 8 Basic stats: COMPLETE Column stats: NONE
value expressions: _col0 (type: bigint)
Reduce Operator Tree:
Group By Operator
aggregations: sum(VALUE._col0)
mode: mergepartial
outputColumnNames: _col0
Statistics: Num rows: 1 Data size: 8 Basic stats: COMPLETE Column stats: NONE
File Output Operator
compressed: true
Statistics: Num rows: 1 Data size: 8 Basic stats: COMPLETE Column stats: NONE
table:
input format: org.apache.hadoop.mapred.SequenceFileInputFormat
output format: org.apache.hadoop.hive.ql.io.HiveSequenceFileOutputFormat
serde: org.apache.hadoop.hive.serde2.lazy.LazySimpleSerDe
Stage: Stage-0
Fetch Operator
limit: -1
Processor Tree:
ListSink
2.2 执行计划最外层
stage dependencies: 各个stage之间的依赖性
stage plan: 各个stage的执行计划
2.2.1 stage dependencies:各个stage之间的依赖性,该部分解读
![](https://img2022.cnblogs.com/blog/2350829/202206/2350829-20220628151013141-189265488.png)
包含两个 (Stage-1、Stage-0) Stage-1 是root Stage。
Stage-0 依赖 Stage-1,说明Stage-1执行完成后执行Stage-0。
2.2.2 stage plan(各个stage的执行计划),该部分解读。执行计划分为Stage-1、Stage-0,Stage-1执行完执行0
![](https://img2022.cnblogs.com/blog/2350829/202206/2350829-20220628151725127-480331756.png)
2.2.2.1 Stage-1中有个Map Reduce,一个MR的执行计划分为两部分
Map Operator Tree: MAP端的执行计划树
Reduce Operator Tree: Reduce端的执行计划树
2.2.2.1.1 Map Operator Tree:MAP端的执行计划树
map端第一个操作肯定是加载表,所以就是 TableScan 表扫描操作
alias: 表名称
Statistics: 表统计信息,包含表中数据条数,数据大小等
Select Operator: 选取操作,常见的属性 :
expressions:需要的字段名称及字段类型
outputColumnNames:输出的列名称
Statistics:表统计信息,包含表中数据条数,数据大小等
Group By Operator:分组聚合操作,常见的属性:
aggregations:显示聚合函数信息
mode:聚合模式,值有 hash:随机聚合,就是hash partition;partial:局部聚合;final:最终聚合
keys:分组的字段,如果没有分组,则没有此字段
outputColumnNames:聚合之后输出列名
Statistics: 表统计信息,包含分组聚合之后的数据条数,数据大小等
Reduce Output Operator:输出到reduce操作
sort order:值为空 不排序; 值为 + 正序排序; 值为 - 倒序排序; 值为 +- 排序的列为两列,第一列为正序,第二列为倒序
Statistics: 统计信息 Num rows: 1 Data size: 8 Basic stats: COMPLETE Column stats: NONE
value expressions: _col0 (type: bigint)
2.2.2.1.2 Reduce Operator Tree: Reduce端的执行计划树
sql是sum,所以算子是group by
Group By Operator
aggregations: 聚合函数信息 sum(VALUE._col0)
mode: mergepartial 聚合模式,值有 hash:随机聚合,就是hash partition;partial:局部聚合;final:最终聚合
outputColumnNames: 输出列名_col0
Statistics: 统计信息Num rows: 1 Data size: 8 Basic stats: COMPLETE Column stats: NONE
File Output Operator
compressed: 是否压缩 true
Statistics: 统计信息Num rows: 1 Data size: 8 Basic stats: COMPLETE Column stats: NONE
table:
input format: 输入格式org.apache.hadoop.mapred.SequenceFileInputFormat
output format: 输出格式org.apache.hadoop.hive.ql.io.HiveSequenceFileOutputFormat
serde: 序列化 org.apache.hadoop.hive.serde2.lazy.LazySimpleSerDe
2.2.2.1.3 其它操作
Filter Operator:过滤操作,常见的属性:
predicate:过滤条件,如sql语句中的where id>=1,则此处显示(id >= 1)
Map Join Operator:join 操作,常见的属性:
condition map:join方式 ,如Inner Join 0 to 1 Left Outer Join0 to 2
keys: join 的条件字段
outputColumnNames: join 完成之后输出的字段
Statistics: join 完成之后生成的数据条数,大小等
File Output Operator:文件输出操作,常见的属性
compressed:是否压缩
table:表的信息,包含输入输出文件格式化方式,序列化方式等
Fetch Operator 客户端获取数据操作,常见的属性:
limit,值为 -1 表示不限制条数,其他值为限制的条数
2.3 子查询性能对比
例1
explain
SELECT
tab1.event_id,
tab2.id
from dw.ods_bdg_db_statistics_tab1 tab1
LEFT JOIN dw.ods_bdg_db_statistics_tab2 tab2
ON tab1.property_id = tab2.id and tab1.dt=tab2.dt
WHERE tab1.dt = '20220627'
and tab1.property_id=983
例2
explain
SELECT tab1.event_id,tab2.id
FROM (select dt,event_id,property_id from dw.ods_bdg_db_statistics_tab1 where dt = '20220627' and property_id=983) tab1
LEFT JOIN dw.ods_bdg_db_statistics_tab2 tab2
ON tab1.property_id = tab2.id and tab1.dt=tab2.dt
WHERE tab1.dt = '20220627'
生成的执行计划,该例子不带子查询性能要好些
left join不带子查询例子,tab2自动带筛选条件性能还高。子查询tab2没筛选,扫描数据量大一些
STAGE DEPENDENCIES:
Stage-1 is a root stage
Stage-0 depends on stages: Stage-1
STAGE PLANS:
Stage: Stage-1
Map Reduce
Map Operator Tree:
TableScan
alias: tab1
Statistics: Num rows: 789 Data size: 82063 Basic stats: COMPLETE Column stats: NONE
Filter Operator
predicate: (property_id = 983) (type: boolean)
Statistics: Num rows: 394 Data size: 40979 Basic stats: COMPLETE Column stats: NONE
Reduce Output Operator
key expressions: property_id (type: int), dt (type: string)
sort order: ++
Map-reduce partition columns: property_id (type: int), dt (type: string)
Statistics: Num rows: 394 Data size: 40979 Basic stats: COMPLETE Column stats: NONE
value expressions: event_id (type: string)
TableScan
alias: tab2
Statistics: Num rows: 7794 Data size: 31177 Basic stats: COMPLETE Column stats: NONE
Filter Operator
predicate: (id = 983) (type: boolean)
Statistics: Num rows: 3897 Data size: 15588 Basic stats: COMPLETE Column stats: NONE
Reduce Output Operator
key expressions: id (type: int), dt (type: string)
sort order: ++
Map-reduce partition columns: id (type: int), dt (type: string)
Statistics: Num rows: 3897 Data size: 15588 Basic stats: COMPLETE Column stats: NONE
Reduce Operator Tree:
Join Operator
condition map:
Left Outer Join0 to 1
keys:
0 property_id (type: int), dt (type: string)
1 id (type: int), dt (type: string)
outputColumnNames: _col1, _col10
Statistics: Num rows: 4286 Data size: 17146 Basic stats: COMPLETE Column stats: NONE
Select Operator
expressions: _col1 (type: string), _col10 (type: int)
outputColumnNames: _col0, _col1
Statistics: Num rows: 4286 Data size: 17146 Basic stats: COMPLETE Column stats: NONE
File Output Operator
compressed: true
Statistics: Num rows: 4286 Data size: 17146 Basic stats: COMPLETE Column stats: NONE
table:
input format: org.apache.hadoop.mapred.SequenceFileInputFormat
output format: org.apache.hadoop.hive.ql.io.HiveSequenceFileOutputFormat
serde: org.apache.hadoop.hive.serde2.lazy.LazySimpleSerDe
Stage: Stage-0
Fetch Operator
limit: -1
Processor Tree:
ListSink