hadoop 数据存储格式

一般而言，存储格式分为列式存储和行存储，对于hadoop来时，列式存储有parquet，rcfile，orcfile等，行存储有SequenceFile,MapFile,Avro Datafile等

hive中的使用
1.orc

 create table test_orc(
 ...
 )
 PARTITIONED BY (day int ) 
 STORED AS ORC 
 LOCATION '/test/test_orc/' 
 tblproperties ("orc.compress"="SNAPPY"); 

默认为tblproperties(“orc.compress”=”ZLIB”);

2.parquet


create table test_parquet(
...
)
PARTITIONED BY (day int )
STORED AS parquet
LOCATION '/test/test_parquet/'
;

CREATE TABLE … STORED AS ORC
ALTER TABLE … SET FILEFORMAT ORC
SET hive.default.fileformat=ORC

参考：
http://blog.csdn.net/bingduanlbd/article/details/52088520
https://www.cnblogs.com/zhenjing/archive/2012/11/02/File-Format.html