优化多个OR查询
我有一个数据库表,我需要验证用户是否输入了相同或部分相同的信息。
这就是我的想法
数据库布局
rec_id (pk), user_id,
name, phone, address_1, address_2, zip,
company, co_phone, co_address_1, co_address_2, co_zip,
billing, bi_phone, bi_address_1, bi_address_2, bi_zip
查询
SELECT rec_id
FROM tbl_name
WHERE user_id = '123456789'
OR '1112223333' IN (phone, co_phone, bi_phone)
OR 'John Doe' IN (name, business, billing)
OR '12345' IN (zip, co_zip, bi_zip)
OR '123 main street' IN (address_1, co_address_1, bi_address_1)
OR 'po box 123' IN (address_2, co_address_2, bi_address_2)
如果任何数据匹配(是的,这将是误报)我需要旧的rec_id。
想知道是否有更好的方法来做到这一点?
谢谢
I have a Database table where I need to validate if a user has entered in the same or partly the same information.
Here is what I'm thinking
The db layout
rec_id (pk), user_id,
name, phone, address_1, address_2, zip,
company, co_phone, co_address_1, co_address_2, co_zip,
billing, bi_phone, bi_address_1, bi_address_2, bi_zip
The Query
SELECT rec_id
FROM tbl_name
WHERE user_id = '123456789'
OR '1112223333' IN (phone, co_phone, bi_phone)
OR 'John Doe' IN (name, business, billing)
OR '12345' IN (zip, co_zip, bi_zip)
OR '123 main street' IN (address_1, co_address_1, bi_address_1)
OR 'po box 123' IN (address_2, co_address_2, bi_address_2)
If any of the data matches (and yes the will be false positives) I need the old rec_id.
Wanted to know if there is a better way to do this?
Thanks
如果你对这篇内容有疑问,欢迎到本站社区发帖提问 参与讨论,获取更多帮助,或者扫码二维码加入 Web 技术交流群。
绑定邮箱获取回复消息
由于您还没有绑定你的真实邮箱,如果其他用户或者作者回复了您的评论,将不能在第一时间通知您!
发布评论
评论(1)
为了使该查询性能良好,您需要为正在测试的每个列建立单独的索引。所有列上的组合索引对于 OR 条件根本没有帮助(但如果有 AND 则会有所帮助)。
但是,我认为无论您添加哪些索引,您的查询都会导致全表扫描。您可能想尝试使用 UNION 而不是 OR 来看看这是否有什么不同:
像这样重写它的想法是现在每个子查询将能够使用索引(当然假设您已经添加了必要的索引)。
For this query to perform well you will need separate indexes for each of the columns you are testing. A combined index on all columns won't help at all for OR conditions (it would help if you had ANDs though).
However I would imagine that your query result in a full table scan regardless of which indexes you add. You might want to try using UNIONs instead of OR to see if that makes a difference:
The idea of rewriting it like this is that now each subquery will be able to use an index (assuming of course that you have added the necessary indexes).