一、hive常用參數

0.常用參數

--@Name:
--@Description: 
--@Type:全量加載
--@Author:---
--@CreateDate:
--@Target:
--@SourceTable:
--@ModifyBy:
--@ModifyDate:
--@ModifyDesc:
--@Copyright 
--設置作業名
set mapred.job.name = hive_xxx(${statisdate});
--Map輸入合並大小
set mapreduce.input.fileinputformat.split.maxsize=
 
300000000;
set mapreduce.input.fileinputformat.split.minsize=100000000;
set mapreduce.input.fileinputformat.split.minsize.per.node=100000000;
set mapreduce.input.fileinputformat.split.minsize.per.rack=100000000;
set hive.input.format = org.apache.hadoop.hive.ql.io.CombineHiveInputFormat;
--設置reduce數目
set hive.exec.reducers.bytes.per.reducer=
 
 300000000;
set hive.exec.reducers.max=300;
--輸出合並
set hive.merge.mapfiles = true;
set hive.merge.mapredfiles = true;
set hive.merge.size.per.task = 128000000;
set hive.merge.smallfiles.avgsize=16000000;
--是否使用mapjoin
set hive.auto.convert.join = false;
--設置默認用戶
use xxx_db;

1.任務名設置

set mapreduce.job.name=
 
xxxx(${statis_date})  # 方便定位具體任務

2.輸入合並參數設置

set mapreduce.input.fileinputformat.split.maxsize=300000000;
set mapreduce.input.fileinputformat.split.minsize=100000000;
set mapreduce.input.fileinputformat.split.minsize.per.node=100000000;
set mapreduce.input.fileinputformat.split.minsize.per.rack=100000000;
set hive.input.format= org.apache.hadoop.hive.ql.io.CombineHiveInputFormat;
set hive.input.format=org.apache.hadoop.hive.ql.io.HiveInputFormat;  --不進行小文件合並

3.輸出合並參數設置

set hive.merge.mapfiles = true    #在Map-only的任務結束時合並小文件
set hive.merge.mapredfiles = true #在Map-Reduce的任務結束時合並小文件
set hive.merge.size.per.task = 256*1000*1000 #合並文件的大小
set hive.merge.smallfiles.avgsize=16000000 #當輸出文件的平均大小小於該值時，啟動一個獨立的map-reduce任務進行文件merge

4.reduce設置

set hive.exec.reducers.bytes.per.reducer= 300000000;
set hive.exec.reducers.max=300;
set mapred.reduce.tasks=10; #固定reduce大小

5.mapjoin參數設置

set hive.auto.convert.join= false; 

6.map端聚合

set hive.map.aggr = true;

7.mapreduce的物理內存、虛擬內存

set mapreduce.map.memory.mb = 4096;
set mapreduce.reduce.memory.mb = 4096;
set mapreduce.map.java.opts=-Xmx3278m;
set mapreduce.reduce.java.opts=-Xmx3278m;
--------------------------------------------------- 
-- set mapreduce.map.memory.mb = 4096;
-- set mapreduce.reduce.memory.mb = 4096;
-- 此參數設計必須在允許範圍內
-- yarn.scheduler.maximum-allocation-mb=8192;
-- yarn.scheduler.minimum-allocation-mb=1024;
--------------------------------------------------- 
-- 堆內存設置要小於物理內存，一般設置為80%
-- set mapreduce.map.java.opts=-Xmx1638m;
-- set mapreduce.reduce.java.opts=-Xmx3278m;
--------------------------------------------------- 
-- Application application_1409135750325_48141 failed 2 times due to AM Container for
-- appattempt_1409135750325_48141_000002 exited with exitCode: 143 due to: Container
-- [pid=4733,containerID=container_1409135750325_48141_02_000001] is running beyond physical memory limits.
-- Current usage: 2.0 GB of 2 GB physical memory used; 6.0 GB of 4.2 GB virtual memory used. Killing container. 
-- #虛擬內存打開：yarn.nodemanager.vmem-check-enabled=true
-- 最大允許使用的虛擬內存=最大可使用的物理內存 * yarn.nodemanager.vmem-pmem-ratio=2.1
-- #物理內存檢查打開：yarn.nodemanager.pmem-check-enabled=true
-- 兩者中有一個超過允許最大內存，此container容器均會被殺
---------------------------------------------------

8.動態分區

set hive.exec.dynamic.partition=true;   
set hive.exec.dynamic.partition.mode=nonstrict;  # 非嚴格模式

9. shuffle端內存溢出oom (BoundedByteArrayOutputStream)

set mapreduce.reduce.shuffle.memory.limit.percent=0.10;

10.map段謂詞下推

set hive.optimize.ppd=true;

11.並行執行

set hive.exec.parallel=true;
set hive.exec.parallel.thread.number=16;   # 並行度

二、hive任務優化

1.分區裁剪

1.查詢涉及分區表時，限制分區範圍
2.使用to_unix_timestamp代替unix_timestamp(),避免全表掃描

2.列裁剪

只讀取查詢中需要用到的列，忽略其他不關心的列
Select * from table_test;
Select field_1,field_2,… from table_test;
Select * 跟select 所有字段是否一樣?(網絡IO，索引)

3.合理設置map、reduce個數

Map數：    splitSize=Math.max(minSize, Math.min(maxSize, blockSize))
reduce數： reducers = Math.min(maxReducers, totalInputFileSize/bytesPerReducer)
# 根據任務運行效率，調整map reduce處理數據量大小

4.group by 優化

set hive.map.aggr=true;
select id,count(1) from test group by id;
set hive.groupby.skewindata = true;
•    先不按GroupBy字段分發，隨機分發做一次聚合
•    額外啟動一輪job，拿前面聚合過的數據按GroupBy字段分發再算結果

5.join優化

大表跟小表之間join時，可打開mapjoin，將小表加載到內存中
set hive.mapjoin.smalltable.filesize   25M
set hive.auto.convert.join = true;
 ps:不能只看文件大小，決定使用使用mapjoin，容易導致OOM(字段、過濾、去重後的記錄數跟文件大小)

--map端join把小表讀入內存
set hive.exec.parallel=true;
select /*+mapjoin(t2)*/
       t1.vendor_cd,
       t2.vendor_cd
from (select vendor_cd
      from tmp_tt
     ) t1
left outer join
     (select vendor_cd
      from tmp_tt
      limit 10
     ) t2
on t1.vendor_cd=t2.vendor_cd
limit 100;

--控制map數，並且用mapjoin實現笛卡爾積
set mapred.reduce.tasks=10;
set hive.input.format=org.apache.hadoop.hive.ql.io.HiveInputFormat;  --不進行小文件合並
set hive.exec.parallel=true;
select /*+mapjoin(t2)*/
       t1.vendor_cd,
       t2.vendor_cd
from (select vendor_cd
      from tmp_tt
      distribute by vendor_cd
     ) t1
left outer join
     (select vendor_cd
      from tmp_tt
      distribute by vendor_cd
     ) t2
limit 100;

6.數據傾斜

--特殊傾斜值的處理（null值很多的時候）
set hive.exec.parallel=true;
select t1.vendor_cd,
       t2.vendor_cd
from (select vendor_cd
      from tmp_tt
     ) t1
left outer join
     (select vendor_cd
      from tmp_tt
     ) t2
on nvl(t1.vendor_cd,concat(‘hive_‘,rand()))=t2.vendor_cd
limit 100;

--當心關聯的類型是否一致,類型不一致可能會導致數據傾斜或者算出意想不到的結果
set hive.exec.parallel=true;
select t1.vendor_cd,
t2.vendor_cd
from (select vendor_cd //int類型
from tmp_tt
) t1
left outer join
(select vendor_cd //string類型
from tmp_tt
) t2
on cast(t1.vendor_cd as string)=t2.vendor_cd
limit 100;

hive參數配置及任務優化