Hive生产脚本常用参数设置[合并小文件，压缩文件，动态分区，数据倾斜...]

合并小文件

hive.merge.mapfile=true;hive.merge.mapredfile=true;hive.merge.size.pertask=32000000; //自定hive.merge.smallfiles.avgsize=16000000; //自定

---

压缩文件

结果压缩

hive.exec.compress.output=true; [同] mapred.output.compress=true;mapred.output.compression.codec=org.apache.hadoop.io.compress.GzipCodec;[通常]mapred.output.compression.type=block;

中间压缩

hive.exec.compress.intermediate=true;hive.intermediate.compression.code=org.apache.hadoop.io.compress.SnappyCodec;[通常]hive.intermediate.compression.type=block;

---

动态分区

hive.mapred.mode=nonstrict;hive.exec.dynamic.partition.mode=nonstrict;hive.exec.max.dynamic.partitions.pernode=10000;[自定，默认100]

---

groupby导致数据倾斜

hive.groupby.skewindata=true;

---

自动mapjoin

hive.auto.convert.join=true;

---

自动本地运行

hive.exec.mode.local.auto=true;hive.exec.mode.local.auto.inputbytes.max=134217728L;[128M]hive.exec.mode.local.auto.tasks.max=2;[自定]

---

对reduce个数限定

mapred.reduce.tasks=-1;[默认不限制]【hive脚本中不起作用，MR中起作用】hive.exec.reduces.bytes.per.reducer=1000000000L;[默认1000M]hive.exec.reducers.max=999;[默认]