I would not recommend using Sqoop to pull data out of HAWQ. It will work but it won't scale so it will be slow when working with any sizable amount of data. The much better way to handle this is to leverage PXF to write data from HAWQ directly to HDFS in parallel. You are also working with SQL so it is will be very easy to transform the data into a format that other tools can consume. Below, I handled the boolean column. --HAWQ create table kirk (
ID INT,
NAME TEXT NOT NULL,
AGE INT NOT NULL,
ADDRESS CHAR(50),
SALARY REAL,
JOIN_DATE DATE,
PRICE NUMERIC,
a BIT(3),
b BIT VARYING(5),
i interval,
ts timestamp,
dec decimal(32,2),
doub double precision,
cv character varying,
t time,
c char(7),
v varchar(20),
isactive boolean
)
distributed randomly; INSERT INTO kirk (ID,NAME,AGE,ADDRESS,SALARY,JOIN_DATE, PRICE, a,b, i, ts,dec,doub,cv,t,c,v, isactive) VALUES
(4, 'Mark', 25, '12 Rich-Mond ', 65000.00, '2007-12-13', 12.44, B'101',B'100', '335','2016-02-22 08:01:23.000',33.3333,0.009,'sdfsadf', '08:01:23.000', 'we ', 'varchar', '0'),
(5, 'David', 27, 'Texas', 85000.00 , '2007-12-13', 45.44,B'101',B'100','775', '2016-02-22 08:01:23.000',33.5555, 0.008, 'white space ','08:01:23.000', 'we', 'varchar ', 'true' ); create writable external table ext_kirk
(
ID INT,
NAME TEXT,
AGE INT,
ADDRESS CHAR(50),
SALARY REAL,
JOIN_DATE DATE,
PRICE NUMERIC,
a BIT(3),
b BIT VARYING(5),
i interval,
ts timestamp,
dec decimal(32,2),
doub double precision,
cv character varying,
t time,
c char(7),
v varchar(20),
isactive varchar(5)
)
location ('pxf://hdb:51200/kirk?Profile=HdfsTextSimple')
format 'text' (delimiter = '|')
distributed randomly; insert into ext_kirk (ID,NAME,AGE,ADDRESS,SALARY,JOIN_DATE, PRICE, a,b, i, ts,dec,doub,cv,t,c,v, isactive)
select ID,NAME,AGE,ADDRESS,SALARY,JOIN_DATE, PRICE, a,b, i, ts,dec,doub,cv,t,c,v, case when isactive then 'TRUE' else 'FALSE' end from kirk;
... View more
