删除主键（Vertica数据库的主键值并不是唯一的）： SELECT ANALYZE_CONSTRAINTS('fb_s.c_log'); 找到key名，再： ALTER TABLE fb_s.c_log DROP CONSTRAINT C_PRIMARY; SELECT ANALYZE_CONSTRAINTS('fb_s.user_info'); ALTER TABLE fb_s.user_info DROP CONST

删除主键（Vertica数据库的主键值并不是唯一的）：
SELECT ANALYZE_CONSTRAINTS('fb_s.c_log');

找到key名，，再：

ALTER TABLE fb_s.c_log DROP CONSTRAINT C_PRIMARY;

SELECT ANALYZE_CONSTRAINTS('fb_s.user_info');

ALTER TABLE fb_s.user_info DROP CONSTRAINT C_PRIMARY;

建用户和SCHEMA ：

CREATE user fb_s_sql IDENTIFIED BY 'password';
CREATE SCHEMA fb_s_sql;

给权限：
GRANT ALL ON SCHEMA fb_s_sql TO fb_s_sql;
GRANT ALL ON SCHEMA fb_s TO fb_s_sql;

GRANT ALL ON TABLE fb_s_sql.sqllog TO fb_s_sql;

建表：

CREATE TABLE fb_s.c_log (
uid int NOT NULL,
cash int,
gold int,
level int,
rtime datetime,
tid varchar(20),
act varchar(50),
item varchar(500),
value int,
value2 int,
time datetime
);

CREATE TABLE fb_s.new_c_log (
uid integer PRIMARY KEY NOT NULL,
cash integer,
gold integer,
level integer,
rtime datetime,
tid varchar(20),
act varchar(50),
item varchar(500),
value integer,
value2 integer,
time datetime NOT NULL
)
PARTITION BY EXTRACT(year FROM time)*100 + EXTRACT(month FROM time);

后一个是按time字段分区

增加及修改字段：

ALTER TABLE fb_s.c_logADD COLUMN value2 integer default 0;
ALTER TABLE fb_s.c_log ALTER COLUMN duration SET DEFAULT 0;
ALTER TABLE fb_s.c_log ALTER COLUMN mesg SET DEFAULT '';


两表之间导数据：
insert into fb_s.c_log (uid,cash,gold,level,rtime,tid,act,item,value,value2,time)
(select * from fb_s.c_logbak);

两库之间导数据：

在源库导出：

vsql -d topcity -U dbadmin -w password -F ',' -At -o fs_user_info.csv -c "SELECT * FROM fb_s.user_info;" &
vsql -d topcity -U dbadmin -w password -F ',' -At -o fs_c_log.csv -c "SELECT * FROM fb_s.c_log;" &

目的库导入：
COPY fb_s.user_info FROM '/opt/fs_user_info.csv' EXCEPTIONS '/tmp/exp.log' DELIMITER ',';

COPY fb_s.c_log FROM '/opt/fs_c_log.csv' EXCEPTIONS '/tmp/exp.log' DELIMITER ',';

PS：

是不是感觉本大神无所不会，想崇拜本神的email到 zhe-jiang.he@hp.com.

帮我在HP明日之星投上一票哦亲！