【Hive学习之八】Hive 调优【重要】
环境
虚拟机:VMware 10
Linux版本:CentOS-6.5-x86_64
客户端:Xshell4
FTP:Xftp4
jdk8
hadoop-3.1.1
apache-hive-3.1.1
一、执行计划
核心思想:把Hive SQL当做Mapreduce程序去优化
以下SQL不会转为Mapreduce来执行
-select仅查询本表字段
-where仅对本表字段做条件过滤
Explain 显示执行计划:EXPLAIN [EXTENDED] query
hive> explain select count(*) from psn2;
OK
Explain
STAGE DEPENDENCIES:
Stage-1 is a root stage
Stage-0 depends on stages: Stage-1 STAGE PLANS:
Stage: Stage-
Map Reduce
Map Operator Tree:
TableScan
alias: psn2
Statistics: Num rows: Data size: Basic stats: COMPLETE Column stats: NONE
Select Operator
Statistics: Num rows: Data size: Basic stats: COMPLETE Column stats: NONE
Group By Operator
aggregations: count()
mode: hash
outputColumnNames: _col0
Statistics: Num rows: Data size: Basic stats: COMPLETE Column stats: NONE
Reduce Output Operator
sort order:
Statistics: Num rows: Data size: Basic stats: COMPLETE Column stats: NONE
value expressions: _col0 (type: bigint)
Execution mode: vectorized
Reduce Operator Tree:
Group By Operator
aggregations: count(VALUE._col0)
mode: mergepartial
outputColumnNames: _col0
Statistics: Num rows: Data size: Basic stats: COMPLETE Column stats: NONE
File Output Operator
compressed: false
Statistics: Num rows: Data size: 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-
Fetch Operator
limit: -
Processor Tree:
ListSink Time taken: 2.7 seconds, Fetched: row(s)
hive>
hive> explain extended select count(*) from psn2;
OK
Explain
STAGE DEPENDENCIES:
Stage- is a root stage
Stage- depends on stages: Stage- STAGE PLANS:
Stage: Stage-
Map Reduce
Map Operator Tree:
TableScan
alias: psn2
Statistics: Num rows: Data size: Basic stats: COMPLETE Column stats: NONE
GatherStats: false
Select Operator
Statistics: Num rows: Data size: Basic stats: COMPLETE Column stats: NONE
Group By Operator
aggregations: count()
mode: hash
outputColumnNames: _col0
Statistics: Num rows: Data size: Basic stats: COMPLETE Column stats: NONE
Reduce Output Operator
null sort order:
sort order:
Statistics: Num rows: Data size: Basic stats: COMPLETE Column stats: NONE
tag: -
value expressions: _col0 (type: bigint)
auto parallelism: false
Execution mode: vectorized
Path -> Alias:
hdfs://PCS102:9820/root/hive_remote/warehouse/psn2/age=10 [psn2]
hdfs://PCS102:9820/root/hive_remote/warehouse/psn2/age=20 [psn2]
Path -> Partition:
hdfs://PCS102:9820/root/hive_remote/warehouse/psn2/age=10
Partition
base file name: age=
input format: org.apache.hadoop.mapred.TextInputFormat
output format: org.apache.hadoop.hive.ql.io.HiveIgnoreKeyTextOutputFormat
partition values:
age
properties:
bucket_count -
collection.delim -
column.name.delimiter ,
columns id,name,likes,address
columns.comments
columns.types int:string:array<string>:map<string,string>
field.delim ,
file.inputformat org.apache.hadoop.mapred.TextInputFormat
file.outputformat org.apache.hadoop.hive.ql.io.HiveIgnoreKeyTextOutputFormat
line.delim location hdfs://PCS102:9820/root/hive_remote/warehouse/psn2/age=10
mapkey.delim :
name default.psn2
numFiles 1
numRows 0
partition_columns age
partition_columns.types int
rawDataSize 0
serialization.ddl struct psn2 { i32 id, string name, list<string> likes, map<string,string> address}
serialization.format ,
serialization.lib org.apache.hadoop.hive.serde2.lazy.LazySimpleSerDe
totalSize 372
transient_lastDdlTime 1548986286
serde: org.apache.hadoop.hive.serde2.lazy.LazySimpleSerDe input format: org.apache.hadoop.mapred.TextInputFormat
output format: org.apache.hadoop.hive.ql.io.HiveIgnoreKeyTextOutputFormat
properties:
bucket_count -
bucketing_version
collection.delim -
column.name.delimiter ,
columns id,name,likes,address
columns.comments
columns.types int:string:array<string>:map<string,string>
field.delim ,
file.inputformat org.apache.hadoop.mapred.TextInputFormat
file.outputformat org.apache.hadoop.hive.ql.io.HiveIgnoreKeyTextOutputFormat
line.delim location hdfs://PCS102:9820/root/hive_remote/warehouse/psn2
mapkey.delim :
name default.psn2
partition_columns age
partition_columns.types int
serialization.ddl struct psn2 { i32 id, string name, list<string> likes, map<string,string> address}
serialization.format ,
serialization.lib org.apache.hadoop.hive.serde2.lazy.LazySimpleSerDe
transient_lastDdlTime
serde: org.apache.hadoop.hive.serde2.lazy.LazySimpleSerDe
name: default.psn2
name: default.psn2
hdfs://PCS102:9820/root/hive_remote/warehouse/psn2/age=20
Partition
base file name: age=
input format: org.apache.hadoop.mapred.TextInputFormat
output format: org.apache.hadoop.hive.ql.io.HiveIgnoreKeyTextOutputFormat
partition values:
age
properties:
bucket_count -
collection.delim -
column.name.delimiter ,
columns id,name,likes,address
columns.comments
columns.types int:string:array<string>:map<string,string>
field.delim ,
file.inputformat org.apache.hadoop.mapred.TextInputFormat
file.outputformat org.apache.hadoop.hive.ql.io.HiveIgnoreKeyTextOutputFormat
line.delim location hdfs://PCS102:9820/root/hive_remote/warehouse/psn2/age=20
mapkey.delim :
name default.psn2
numFiles
numRows
partition_columns age
partition_columns.types int
rawDataSize
serialization.ddl struct psn2 { i32 id, string name, list<string> likes, map<string,string> address}
serialization.format ,
serialization.lib org.apache.hadoop.hive.serde2.lazy.LazySimpleSerDe
totalSize
transient_lastDdlTime
serde: org.apache.hadoop.hive.serde2.lazy.LazySimpleSerDe input format: org.apache.hadoop.mapred.TextInputFormat
output format: org.apache.hadoop.hive.ql.io.HiveIgnoreKeyTextOutputFormat
properties:
bucket_count -
bucketing_version
collection.delim -
column.name.delimiter ,
columns id,name,likes,address
columns.comments
columns.types int:string:array<string>:map<string,string>
field.delim ,
file.inputformat org.apache.hadoop.mapred.TextInputFormat
file.outputformat org.apache.hadoop.hive.ql.io.HiveIgnoreKeyTextOutputFormat
line.delim location hdfs://PCS102:9820/root/hive_remote/warehouse/psn2
mapkey.delim :
name default.psn2
partition_columns age
partition_columns.types int
serialization.ddl struct psn2 { i32 id, string name, list<string> likes, map<string,string> address}
serialization.format ,
serialization.lib org.apache.hadoop.hive.serde2.lazy.LazySimpleSerDe
transient_lastDdlTime
serde: org.apache.hadoop.hive.serde2.lazy.LazySimpleSerDe
name: default.psn2
name: default.psn2
Truncated Path -> Alias:
/psn2/age= [psn2]
/psn2/age= [psn2]
Needs Tagging: false
Reduce Operator Tree:
Group By Operator
aggregations: count(VALUE._col0)
mode: mergepartial
outputColumnNames: _col0
Statistics: Num rows: Data size: Basic stats: COMPLETE Column stats: NONE
File Output Operator
compressed: false
GlobalTableId:
directory: hdfs://PCS102:9820/tmp/hive/root/6f8ff71f-87bd-4d46-9f9a-516708d65459/hive_2019-02-19_10-58-42_159_2637812497308639143-1/-mr-10001/.hive-staging_hive_2019-02-19_10-58-42_159_2637812497308639143-1/-ext-10002
NumFilesPerFileSink:
Statistics: Num rows: Data size: Basic stats: COMPLETE Column stats: NONE
Stats Publishing Key Prefix: hdfs://PCS102:9820/tmp/hive/root/6f8ff71f-87bd-4d46-9f9a-516708d65459/hive_2019-02-19_10-58-42_159_2637812497308639143-1/-mr-10001/.hive-staging_hive_2019-02-19_10-58-42_159_2637812497308639143-1/-ext-10002/
table:
input format: org.apache.hadoop.mapred.SequenceFileInputFormat
output format: org.apache.hadoop.hive.ql.io.HiveSequenceFileOutputFormat
properties:
columns _col0
columns.types bigint
escape.delim \
hive.serialization.extend.additional.nesting.levels true
serialization.escape.crlf true
serialization.format
serialization.lib org.apache.hadoop.hive.serde2.lazy.LazySimpleSerDe
serde: org.apache.hadoop.hive.serde2.lazy.LazySimpleSerDe
TotalFiles:
GatherStats: false
MultiFileSpray: false Stage: Stage-
Fetch Operator
limit: -
Processor Tree:
ListSink Time taken: 0.142 seconds, Fetched: row(s)
hive>
二、运行模式
(1)分为 本地模式 和 集群模式
(2)开启本地模式(对于数据量少的表情况):
set hive.exec.mode.local.auto=true;
注意:
hive.exec.mode.local.auto.inputbytes.max默认值为128M
表示加载文件的最大值,若大于该配置仍会以集群方式来运行!
hive> set hive.exec.mode.local.auto=true;
hive> select count(*) from psn21;
Automatically selecting local only mode for query
Query ID = root_20190219144810_0bafff9e-1c40-45f6-b687-60c5d13c9f0c
Total jobs =
Launching Job out of
Number of reduce tasks determined at compile time:
In order to change the average load for a reducer (in bytes):
set hive.exec.reducers.bytes.per.reducer=<number>
In order to limit the maximum number of reducers:
set hive.exec.reducers.max=<number>
In order to set a constant number of reducers:
set mapreduce.job.reduces=<number>
Job running in-process (local Hadoop)
-- ::, Stage- map = %, reduce = %
Ended Job = job_local1827024396_0002
MapReduce Jobs Launched:
Stage-Stage-: HDFS Read: HDFS Write: SUCCESS
Total MapReduce CPU Time Spent: msec
OK
_c0 Time taken: 1.376 seconds, Fetched: row(s)
hive> set hive.exec.mode.local.auto=false;
hive> select count(*) from psn21;
Query ID = root_20190219144841_6fd11106-5db1--8b0b-884697b558df
Total jobs =
Launching Job out of
Number of reduce tasks determined at compile time:
In order to change the average load for a reducer (in bytes):
set hive.exec.reducers.bytes.per.reducer=<number>
In order to limit the maximum number of reducers:
set hive.exec.reducers.max=<number>
In order to set a constant number of reducers:
set mapreduce.job.reduces=<number>
Starting Job = job_1548397153910_0013, Tracking URL = http://PCS102:8088/proxy/application_1548397153910_0013/
Kill Command = /usr/local/hadoop-3.1./bin/mapred job -kill job_1548397153910_0013
Hadoop job information for Stage-: number of mappers: ; number of reducers:
-- ::, Stage- map = %, reduce = %
-- ::, Stage- map = %, reduce = %, Cumulative CPU 2.87 sec
-- ::, Stage- map = %, reduce = %, Cumulative CPU 6.28 sec
MapReduce Total cumulative CPU time: seconds msec
Ended Job = job_1548397153910_0013
MapReduce Jobs Launched:
Stage-Stage-: Map: Reduce: Cumulative CPU: 6.28 sec HDFS Read: HDFS Write: SUCCESS
Total MapReduce CPU Time Spent: seconds msec
OK
_c0 Time taken: 18.923 seconds, Fetched: row(s)
hive>
三、并行计算
通过设置以下参数开启并行模式(需要关闭本地模式):
set hive.exec.parallel=true;
另:hive.exec.parallel.thread.number:一次SQL计算中允许并行执行的job个数的最大值
hive> set hive.exec.parallel;
hive.exec.parallel=false
hive> select t1.cnt1,t2.cnt2 from
> (select count(id) cnt1 from psn21) t1,
> (select count(name) cnt2 from psn21)t2;
Warning: Map Join MAPJOIN[][bigTable=?] in task 'Stage-4:MAPRED' is a cross product
Warning: Map Join MAPJOIN[][bigTable=?] in task 'Stage-5:MAPRED' is a cross product
Warning: Shuffle Join JOIN[][tables = [$hdt$_0, $hdt$_1]] in Stage 'Stage-2:MAPRED' is a cross product
Query ID = root_20190219145608_b4f3d4e9-b858-41be-9ddc-6eccff0ec9d9
Total jobs =
Launching Job 1 out of 5
Number of reduce tasks determined at compile time:
In order to change the average load for a reducer (in bytes):
set hive.exec.reducers.bytes.per.reducer=<number>
In order to limit the maximum number of reducers:
set hive.exec.reducers.max=<number>
In order to set a constant number of reducers:
set mapreduce.job.reduces=<number>
Starting Job = job_1548397153910_0014, Tracking URL = http://PCS102:8088/proxy/application_1548397153910_0014/
Kill Command = /usr/local/hadoop-3.1./bin/mapred job -kill job_1548397153910_0014
Hadoop job information for Stage-: number of mappers: ; number of reducers:
-- ::, Stage- map = %, reduce = %
-- ::, Stage- map = %, reduce = %, Cumulative CPU 2.85 sec
-- ::, Stage- map = %, reduce = %, Cumulative CPU 6.04 sec
MapReduce Total cumulative CPU time: seconds msec
Ended Job = job_1548397153910_0014
Launching Job 2 out of 5
Number of reduce tasks determined at compile time:
In order to change the average load for a reducer (in bytes):
set hive.exec.reducers.bytes.per.reducer=<number>
In order to limit the maximum number of reducers:
set hive.exec.reducers.max=<number>
In order to set a constant number of reducers:
set mapreduce.job.reduces=<number>
Starting Job = job_1548397153910_0015, Tracking URL = http://PCS102:8088/proxy/application_1548397153910_0015/
Kill Command = /usr/local/hadoop-3.1./bin/mapred job -kill job_1548397153910_0015
Hadoop job information for Stage-: number of mappers: ; number of reducers:
-- ::, Stage- map = %, reduce = %
-- ::, Stage- map = %, reduce = %, Cumulative CPU 2.8 sec
-- ::, Stage- map = %, reduce = %, Cumulative CPU 5.92 sec
MapReduce Total cumulative CPU time: seconds msec
Ended Job = job_1548397153910_0015
Stage- is selected by condition resolver.
Stage- is filtered out by condition resolver.
Stage- is filtered out by condition resolver.
SLF4J: Found binding in [jar:file:/usr/local/apache-hive-3.1.-bin/lib/log4j-slf4j-impl-2.10..jar!/org/slf4j/impl/StaticLoggerBinder.class]
SLF4J: Actual binding is of type [org.apache.logging.slf4j.Log4jLoggerFactory]
-- :: Starting to launch local task to process map join; maximum memory =
-- :: Dump the side-table for tag: with group count: into file: file:/tmp/root/6f8ff71f-87bd-4d46-9f9a-516708d65459/hive_2019--19_14--08_997_6748376838876035123-/-local-/HashTable-Stage-/MapJoin-mapfile01--.hashtable2019-- :: Uploaded File to: file:/tmp/root/6f8ff71f-87bd-4d46-9f9a-516708d65459/hive_2019--19_14--08_997_6748376838876035123-/-local-/HashTable-Stage-/MapJoin-mapfile01--.hashtable ( bytes) Execution completed successfully
MapredLocal task succeeded
Launching Job 4 out of 5
Number of reduce tasks is set to since there's no reduce operator
Starting Job = job_1548397153910_0016, Tracking URL = http://PCS102:8088/proxy/application_1548397153910_0016/
Kill Command = /usr/local/hadoop-3.1./bin/mapred job -kill job_1548397153910_0016
Hadoop job information for Stage-: number of mappers: ; number of reducers:
-- ::, Stage- map = %, reduce = %
-- ::, Stage- map = %, reduce = %, Cumulative CPU 3.07 sec
MapReduce Total cumulative CPU time: seconds msec
Ended Job = job_1548397153910_0016
MapReduce Jobs Launched:
Stage-Stage-: Map: Reduce: Cumulative CPU: 6.04 sec HDFS Read: HDFS Write: SUCCESS
Stage-Stage-: Map: Reduce: Cumulative CPU: 5.92 sec HDFS Read: HDFS Write: SUCCESS
Stage-Stage-: Map: Cumulative CPU: 3.07 sec HDFS Read: HDFS Write: SUCCESS
Total MapReduce CPU Time Spent: seconds msec
OK
t1.cnt1 t2.cnt2 Time taken: 59.527 seconds, Fetched: row(s)
hive> set hive.exec.parallel=true;
hive> (select count(name) cnt2 from psn21)t2;
FAILED: ParseException line : extraneous input 't2' expecting EOF near '<EOF>'
hive> select t1.cnt1,t2.cnt2 from
> (select count(id) cnt1 from psn21) t1,
> (select count(name) cnt2 from psn21)t2;
Warning: Map Join MAPJOIN[][bigTable=?] in task 'Stage-4:MAPRED' is a cross product
Warning: Map Join MAPJOIN[][bigTable=?] in task 'Stage-5:MAPRED' is a cross product
Warning: Shuffle Join JOIN[][tables = [$hdt$_0, $hdt$_1]] in Stage 'Stage-2:MAPRED' is a cross product
Query ID = root_20190219145918_2f98437b--41a4-905b-4c6a3a160d46
Total jobs =
Launching Job 1 out of 5
Launching Job 2 out of 5
Number of reduce tasks determined at compile time:
In order to change the average load for a reducer (in bytes):
set hive.exec.reducers.bytes.per.reducer=<number>
In order to limit the maximum number of reducers:
set hive.exec.reducers.max=<number>
In order to set a constant number of reducers:
set mapreduce.job.reduces=<number>
Number of reduce tasks determined at compile time:
In order to change the average load for a reducer (in bytes):
set hive.exec.reducers.bytes.per.reducer=<number>
In order to limit the maximum number of reducers:
set hive.exec.reducers.max=<number>
In order to set a constant number of reducers:
set mapreduce.job.reduces=<number>
Starting Job = job_1548397153910_0018, Tracking URL = http://PCS102:8088/proxy/application_1548397153910_0018/
Kill Command = /usr/local/hadoop-3.1./bin/mapred job -kill job_1548397153910_0018
Starting Job = job_1548397153910_0017, Tracking URL = http://PCS102:8088/proxy/application_1548397153910_0017/
Kill Command = /usr/local/hadoop-3.1./bin/mapred job -kill job_1548397153910_0017
Hadoop job information for Stage-: number of mappers: ; number of reducers:
-- ::, Stage- map = %, reduce = %
-- ::, Stage- map = %, reduce = %, Cumulative CPU 3.0 sec
-- ::, Stage- map = %, reduce = %, Cumulative CPU 6.25 sec
MapReduce Total cumulative CPU time: seconds msec
Ended Job = job_1548397153910_0018
Hadoop job information for Stage-: number of mappers: ; number of reducers:
-- ::, Stage- map = %, reduce = %
-- ::, Stage- map = %, reduce = %, Cumulative CPU 2.74 sec
-- ::, Stage- map = %, reduce = %, Cumulative CPU 5.9 sec
MapReduce Total cumulative CPU time: seconds msec
Ended Job = job_1548397153910_0017
Stage- is selected by condition resolver.
Stage- is filtered out by condition resolver.
Stage- is filtered out by condition resolver.
SLF4J: Found binding in [jar:file:/usr/local/apache-hive-3.1.-bin/lib/log4j-slf4j-impl-2.10..jar!/org/slf4j/impl/StaticLoggerBinder.class]SLF4J: Actual binding is of type [org.apache.logging.slf4j.Log4jLoggerFactory]
-- :: Dump the side-table for tag: with group count: into file: file:/tmp/root/6f8ff71f-87bd-4d46-9f9a-516708d65459/hive_2019--19_14--18_586_8660726948780795909-/-local-/HashTable-Stage-/MapJoin-mapfile21--.hashtable2019-- :: Uploaded File to: file:/tmp/root/6f8ff71f-87bd-4d46-9f9a-516708d65459/hive_2019--19_14--18_586_8660726948780795909-/-local-/HashTable-Stage-/MapJoin-mapfile21--.hashtable ( bytes) Execution completed successfully
MapredLocal task succeeded
Launching Job 4 out of 5
Number of reduce tasks is set to since there's no reduce operator
Starting Job = job_1548397153910_0019, Tracking URL = http://PCS102:8088/proxy/application_1548397153910_0019/
Kill Command = /usr/local/hadoop-3.1./bin/mapred job -kill job_1548397153910_0019
Hadoop job information for Stage-: number of mappers: ; number of reducers:
-- ::, Stage- map = %, reduce = %
-- ::, Stage- map = %, reduce = %, Cumulative CPU 2.95 sec
MapReduce Total cumulative CPU time: seconds msec
Ended Job = job_1548397153910_0019
MapReduce Jobs Launched:
Stage-Stage-: Map: Reduce: Cumulative CPU: 6.25 sec HDFS Read: HDFS Write: SUCCESS
Stage-Stage-: Map: Reduce: Cumulative CPU: 5.9 sec HDFS Read: HDFS Write: SUCCESS
Stage-Stage-: Map: Cumulative CPU: 2.95 sec HDFS Read: HDFS Write: SUCCESS
Total MapReduce CPU Time Spent: seconds msec
OK
t1.cnt1 t2.cnt2 Time taken: 64.206 seconds, Fetched: row(s)
hive>
四、严格模式
通过设置以下参数开启严格模式:
set hive.mapred.mode=strict;
(默认为:nonstrict非严格模式)
查询限制:
1、对于分区表,必须添加where对于分区字段的条件过滤;
hive> set hive.mapred.mode=nonstrict;
hive> select * from psn22;
OK
psn22.id psn22.name psn22.likes psn22.address psn22.age psn22.sex
小明1 ["lol","book","movie"] {"beijing":"shangxuetang","shanghai":"pudong"} boy
小明2 ["lol","book","movie"] {"beijing":"shangxuetang","shanghai":"pudong"} man
小明5 ["lol","book","movie"] {"beijing":"shangxuetang","shanghai":"pudong"} boy
小明3 ["lol","book","movie"] {"beijing":"shangxuetang","shanghai":"pudong"} boy
小明6 ["lol","book","movie"] {"beijing":"shangxuetang","shanghai":"pudong"} man
小明4 ["lol","book","movie"] {"beijing":"shangxuetang","shanghai":"pudong"} man
Time taken: 0.186 seconds, Fetched: row(s)
hive> set hive.mapred.mode=strict;
hive> select * from psn22;
FAILED: SemanticException [Error ]: Queries against partitioned tables without a partition filter are disabled for safety reasons. If you know what you are doing, please set hive.strict.checks.no.partition.filter to false and make sure that hive.mapred.mode is not set to 'strict' to proceed. Note that you may get errors or incorrect results if you make a mistake while using some of the unsafe features. No partition predicate for Alias "psn22" Table "psn22"
hive> select * from psn22 where age= and sex='boy';
OK
psn22.id psn22.name psn22.likes psn22.address psn22.age psn22.sex
小明1 ["lol","book","movie"] {"beijing":"shangxuetang","shanghai":"pudong"} boy
Time taken: 0.282 seconds, Fetched: row(s)
hive>
2、order by语句必须包含limit输出限制
hive> set hive.mapred.mode=strict;
hive> select * from psn21 order by id;
FAILED: SemanticException : Order by-s without limit are disabled for safety reasons. If you know what you are doing, please set hive.strict.checks.orderby.no.limit to false and make sure that hive.mapred.mode is not set to 'strict' to proceed. Note that you may get errors or incorrect results if you make a mistake while using some of the unsafe features.. Error encountered near token 'id'
hive> select * from psn21 order by id limit ;
Automatically selecting local only mode for query
Query ID = root_20190219143842_b465a76f-a890-4bdc-aa76-b713c3ea13c0
Total jobs =
Launching Job out of
Number of reduce tasks determined at compile time:
In order to change the average load for a reducer (in bytes):
set hive.exec.reducers.bytes.per.reducer=<number>
In order to limit the maximum number of reducers:
set hive.exec.reducers.max=<number>
In order to set a constant number of reducers:
set mapreduce.job.reduces=<number>
Job running in-process (local Hadoop)
-- ::, Stage- map = %, reduce = %
Ended Job = job_local1585589360_0001
MapReduce Jobs Launched:
Stage-Stage-: HDFS Read: HDFS Write: SUCCESS
Total MapReduce CPU Time Spent: msec
OK
psn21.id psn21.name psn21.age psn21.sex psn21.likes psn21.address
小明1 boy ["lol","book","movie"] {"beijing":"shangxuetang","shanghai":"pudong"}
小明2 man ["lol","book","movie"] {"beijing":"shangxuetang","shanghai":"pudong"}
Time taken: 1.89 seconds, Fetched: row(s)
hive>
3、限制执行笛卡尔积的查询。
五、Hive排序
-Order By - 对于查询结果做全排序,只允许有一个reduce处理
(当数据量较大时,应慎用。严格模式下,必须结合limit来使用)
-Sort By - 对于单个reduce的数据进行排序
-Distribute By - 分区排序,经常和Sort By结合使用
-Cluster By - 相当于 Sort By + Distribute By
(Cluster By不能通过asc、desc的方式指定排序规则;
可通过 distribute by column sort by column asc|desc 的方式)
六、Hive Join(去掉MapReduce中的shuffle)
Join计算时,将小表(驱动表)放在join的左边
Map Join:在Map端完成Join
两种实现方式:
1、SQL方式,在SQL语句中添加MapJoin标记(mapjoin hint)
语法:
SELECT /*+ MAPJOIN(smallTable) */ smallTable.key, bigTable.value
FROM smallTable JOIN bigTable ON smallTable.key = bigTable.key;
2、开启自动的MapJoin
通过修改以下配置启用自动的mapjoin:
set hive.auto.convert.join = true;
(该参数为true时,Hive自动对左边的表统计量,如果是小表就加入内存,即对小表使用Map join)
相关配置参数:
hive.mapjoin.smalltable.filesize;
(大表小表判断的阈值,如果表的大小小于该值则会被加载到内存中运行)
hive.ignore.mapjoin.hint;
(默认值:true;是否忽略mapjoin hint 即mapjoin标记)
hive.auto.convert.join.noconditionaltask;
(默认值:true;将普通的join转化为普通的mapjoin时,是否将多个mapjoin转化为一个mapjoin)
hive.auto.convert.join.noconditionaltask.size;
(将多个mapjoin转化为一个mapjoin时,其表的最大值)
七、Map-Side聚合(相当于MapReduce中的combine聚合)
通过设置以下参数开启在Map端的聚合:
set hive.map.aggr=true;
相关配置参数:
hive.groupby.mapaggr.checkinterval:
map端group by执行聚合时处理的多少行数据(默认:100000)
hive.map.aggr.hash.min.reduction:
进行聚合的最小比例(预先对100000条数据做聚合,若聚合之后的数据量/100000的值大于该配置0.5,则不会聚合)
hive.map.aggr.hash.percentmemory:
map端聚合使用的内存的最大值
hive.map.aggr.hash.force.flush.memory.threshold:
map端做聚合操作是hash表的最大可用内容,大于该值则会触发flush
hive.groupby.skewindata
是否对GroupBy产生的数据倾斜做优化,默认为false
八、控制Hive中Map以及Reduce的数量
Map数量相关的参数
mapred.max.split.size
一个split的最大值,即每个map处理文件的最大值
mapred.min.split.size.per.node
一个节点上split的最小值
mapred.min.split.size.per.rack
一个机架上split的最小值
Reduce数量相关的参数:
mapred.reduce.tasks
强制指定reduce任务的数量
hive.exec.reducers.bytes.per.reducer
每个reduce任务处理的数据量
hive.exec.reducers.max
每个任务最大的reduce数
九、Hive - JVM重用
适用场景:
1、小文件个数过多
2、task个数过多
通过 set mapred.job.reuse.jvm.num.tasks=n; 来设置(n为task插槽个数)
缺点:设置开启之后,task插槽会一直占用资源,不论是否有task运行,
直到所有的task即整个job全部执行完成时,才会释放所有的task插槽资源!
【Hive学习之八】Hive 调优【重要】的更多相关文章
- Spark学习之Spark调优与调试(7)
Spark学习之Spark调优与调试(7) 1. 对Spark进行调优与调试通常需要修改Spark应用运行时配置的选项. 当创建一个SparkContext时就会创建一个SparkConf实例. 2. ...
- 大数据技术之_08_Hive学习_04_压缩和存储(Hive高级)+ 企业级调优(Hive优化)
第8章 压缩和存储(Hive高级)8.1 Hadoop源码编译支持Snappy压缩8.1.1 资源准备8.1.2 jar包安装8.1.3 编译源码8.2 Hadoop压缩配置8.2.1 MR支持的压缩 ...
- hive优化之参数调优
1.hive参数优化之默认启用本地模式 启动hive本地模式参数,一般建议将其设置为true,即时刻启用: hive (chavin)> set hive.exec.mode.local.aut ...
- Hive Tuning(五) 标准调优清单
Hive的标准调优清单,我们可以对照着来做我们的查询优化!
- Hive(十二)【调优】
目录 1.Fetch抓取 2.本地模式 3.表的优化 3.1大小表join 3.2大表Join大表 3.3map join 3.4group By 3.5 count(distinct) 3.6笛卡尔 ...
- Spark学习之Spark调优与调试(二)
下面来看看更复杂的情况,比如,当调度器进行流水线执行(pipelining),或把多个 RDD 合并到一个步骤中时.当RDD 不需要混洗数据就可以从父节点计算出来时,调度器就会自动进行流水线执行.上一 ...
- Spark学习之Spark调优与调试(一)
一.使用SparkConf配置Spark 对 Spark 进行性能调优,通常就是修改 Spark 应用的运行时配置选项.Spark 中最主要的配置机制是通过 SparkConf 类对 Spark 进行 ...
- hive学习(二) hive操作
hive ddl 操作官方手册https://cwiki.apache.org/confluence/display/Hive/LanguageManual+DDL hive dml 操作官方手 ...
- hive sql的参数调优
shuffle优化之减少shuffle数据量 1.谓词下推 hive.optimize.ppd ,默认为true. 所谓谓词下推就是过滤条件如果写在shuffle操作后面,就提前过滤掉,减少参与sh ...
随机推荐
- @media screen and (max-width: 960px)与@media (max-width: 960px) 有screen与没有screen的区别
我们先来看下下面这段代码,估计很多人在响应式的网站CSS很经常看到类似下面的这段代码: @media screen and (max-width: 960px){ body{ background: ...
- 命令行安装kvm虚拟机、桥接网络、用virt-manager管理
宿主机CentOS Linux release 7.2.1511 (Core),内核3.10.0-327.el7.x86_64 1.配置宿主机网络桥接 想让虚拟机有自己的ip且外网可访问,需要在安装虚 ...
- Word Embedding理解
一直以来感觉好多地方都吧Word Embedding和word2vec混起来一起说,所以导致对这俩的区别不是很清楚. 其实简单说来就是word embedding包含了word2vec,word2ve ...
- FastDFS的使用
1.FastDFS 1.1. 什么是FastDFS? FastDFS是用c语言编写的一款开源的分布式文件系统.FastDFS为互联网量身定制,充分考虑了冗余备份.负载均衡.线性扩容等机制,并注重高可用 ...
- 【托业】【新托业TOEIC新题型真题】学习笔记9-题库七+八--P4-5
109.intend 意为“打算,意欲”,含有将来的含义,故不用将来时态 110.must do sth 必须做某事 111.recession 经济衰退,不景气 rebound 反弹:反应 recr ...
- Linux查看访问IP
Linux查看访问IP https://blog.csdn.net/tojohnonly/article/details/82772323
- MongoDB pymongo模块 查询
查询 mongo_db 类似于 服务器命令行的db 我们可以db.user.find() 查询 find() 需要加上列表 import pymongo mongo_client = pymongo. ...
- 启动Jmeter4.0 后弹出命令窗口提示信息是什么意思?
启动Jmeter4.0 后弹出命令窗口提示信息: =========================================================================== ...
- spark提交任务的三种的方法
在学习Spark过程中,资料中介绍的提交Spark Job的方式主要有三种: 第一种: 通过命令行的方式提交Job,使用spark 自带的spark-submit工具提交,官网和大多数参考资料都是已这 ...
- Python复习基础篇
4.4 使用列表的一部分 Python切片(切片就是取值的呗) print(players[0:3]) 中括号,冒号隔开,顾头不顾尾. print([:4]) 从开始取还是会顾尾的 pri ...