select distinct brands.*from product_prices
inner join products on products.id = product_prices.product_id
inner join brands on brands.id = products.brand_id
inner join product_batches on product_prices.batch_id = product_batches.id
where (quotation_batches.id =289859 OR quotation_batches.group_id =289859)
and products.publish_stage IN (0,1)
and brands.status NOT IN (1,2)
由于三张表都比较大，顺序的影响可以先不考虑；
select * from
( select distinct * from products where products.publish_stage IN (0,1) ) t1
inner join ( select distinct brands.*from brands where brands.status NOT IN (1,2) ) t2
on t1.batch_id = t2.id
inner join product_prices t3 on t1.id = t3.product_id
inner join product_batches t4 on t3.batch_id = t4.id
我的思路：尽量缩小联查的每个表的数据量；
下面在网上找到关于：Ruby通过OCI8操作Oracle存储的事例
1.读blob的存储过程
CREATE OR REPLACE PROCEDURE "P_READ_IMAGE"
(
V_IMG_ID IN NUMBER,
CUR_RESULT OUT SYS_REFCURSOR
)
AS
BEGIN
OPEN CUR_RESULT FOR
SELECT ID,IMG
FROM T_IMAGE
WHERE ID=V_IMG_ID ;
END;
5、ruby读图片操作
require 'oci8'
h_conn = OCI8.new(DB_USER, DB_PASSWORD, DB_SERVER)
s_photo_target_path = "~/222.jpg"
photo_id = 1
begin
cursor = h_conn.parse('begin P_READ_IMAGE(
:img_id,
:list
); end;')
cursor.bind_param(':img_id', photo_id)
cursor.bind_param(':list', nil, OCI8::Cursor)
cursor.exec()
ret_cursor = cursor[':list']
puts ret_cursor.getColNames.join(",")
while row = ret_cursor.fetch()
puts row[0]
File.open(s_photo_target_path, 'wb') do |f|
f.write(row[1].read)
end
break;
end
rescue OCIError
puts '-'*80
puts "Code: " + $!.code.to_s
puts "Desc: " + $!.message
puts '-'*80
end
