Hive常用命令及设置


alter table ad_app.app_accounting_daily add columns (return_cost bigint) cascade;


数据倾斜

set hive.map.aggr=true;

set hive.groupby.skewindata=true;


set mapreduce.job.queuename = root.data.adonline;


set mapred.reduce.tasks=8192;

set mapreduce.job.priority=VERY_HIGH;

Set hive.groupby.mapaggr.checkinterval = 100000


set mapred.output.compression.codec=org.apache.hadoop.io.compress.GzipCodec;

set mapred.map.output.compression.codec=org.apache.hadoop.io.compress.GzipCodec;

set hive.exec.compress.intermediate=true;

set mapred.compress.map.output=true;


set mapred.max.split.size=256000000;        -- 决定每个map处理的最大的文件大小，单位为B

set mapred.min.split.size.per.node=1;         -- 节点中可以处理的最小的文件大小

set mapred.min.split.size.per.rack=1;    



控制map数量

set mapred.max.split.size=128000000;

set mapred.min.split.size.per.node=64000000;

set mapred.min.split.size.per.rack=64000000;

set hive.input.format=org.apache.hadoop.hive.ql.io.CombineHiveInputFormat;


设置reducer最大个数

set hive.exec.reducers.max=8192;



控制map reduce内存

set mapreduce.map.memory.mb=4096;

set mapreduce.reduce.memory.mb=4096;

set mapreduce.reduce.java.opts=-Xmx3800M;

set mapred.child.map.java.opts=-Xmx3800M;

set mapreduce.map.java.opts=-Xmx3800M;


ALTER TABLE table_name ADD IF NOT EXISTS PARTITION (dt='20130101') LOCATION '/user/hadoop/warehouse/table_name/dt=20130101'; //一次添加一个分区