hive创建分区,hive创建


 HIVE的分区通过在创建表时启用partitionby实现,用来partition的维度并不是实际数据的某一列,具体分区的标志是由插入内容时给定的。当要查询某一分区的内容时可以采用where语句,形似where tablename.partition_key >a来实现。 创建含分区的表。 命令原型: CREATE TABLE page_view(viewTime INT, userid BIGINT,      page_urlSTRING, referrer_url STRING,      ip STRINGCOMMENT 'IP Address of the User') COMMENT 'This isthe page view table' PARTITIONED BY(dtSTRING, country STRING) CLUSTEREDBY(userid) SORTED BY(viewTime) INTO 32 BUCKETS ROW FORMATDELIMITED    FIELDSTERMINATED BY '\001'    COLLECTION ITEMSTERMINATED BY '\002'    MAP KEYSTERMINATED BY '\003' STORED ASSEQUENCEFILE;
Eg: 建表: CREATE TABLE c02_clickstat_fatdt1 (yyyymmdd  string, id              INT, ip               string, country          string, cookie_id        string, page_id          string  ,   clickstat_url_id int, query_string     string, refer            string )PARTITIONED BY(dt STRING) row format delimited fields terminated by '\005' stored astextfile; 装载数据: LOAD DATA INPATH'/user/admin/SqlldrDat/CnClickstat/20101101/19/clickstat_gp_fatdt0/0' OVERWRITEINTO TABLE c02_clickstat_fatdt1 PARTITION(dt='20101101'); 访问某一个分区 SELECT count(*)     FROMc02_clickstat_fatdt1 a     WHERE a.dt >='20101101' AND a.dt < '20101102';

相关内容