[英]Performance: WHERE IN clause vs (INSERT + INNER JOIN)
I have a use-case where I need to perform a very high number of SELECT SQL 我有一个用例,需要执行大量的SELECT SQL
I have two approaches at this moment: 目前,我有两种方法:
Query by a list of identifiers. 通过标识符列表进行查询。 So, I first used WHERE IN clause:
因此,我首先使用了WHERE IN子句:
I can create a table, let's say, CACHE_TABLE, and first INSERT the identifiers ( 1,2,3,8,11,78,59,65,74,25,36,54558,78854,558 ) into it by a unique key CACHEID and the JOIN this CACHE_TABLE with MAIN_TABLE to get the desired result: 我可以创建一个表,比如说CACHE_TABLE,然后首先将一个唯一的标识符(1,2,3,8,11,78,59,65,74,25,36,54558,78854,558)插入其中键CACHEID并将此CACHE_TABLE与MAIN_TABLE联接以获得所需结果:
Performance is really critical in my use-case. 在我的用例中,性能确实至关重要。 So I wanted to know if the approach #2 would yield better performance than approach #1.
所以我想知道方法2是否会比方法1产生更好的性能。 Also, if there is a better alternative approach(s) for this
另外,如果有更好的替代方法
Thanks a ton in Advance!! 在此先感谢一吨!
your answer is best performance approach #2. 您的答案是最佳性能方法2。 In my experience IN is a very slow operator, since SQL normally evaluates it as a series of WHERE clauses separated by "OR" (WHERE x=Y OR x=Z OR...).
以我的经验,IN是一个非常慢的运算符,因为SQL通常将其评估为一系列由“ OR”(WHERE x = Y OR x = Z OR ...)分隔的WHERE子句。 As with ALL THINGS SQL though, your mileage may vary.
与ALL THINGS SQL一样,您的里程可能会有所不同。 The speed will depend a lot on indexes
速度将在很大程度上取决于索引
You need to test the two approaches. 您需要测试这两种方法。
For a single query, I would expect in
to win in most cases -- simply because creating the table and then uses it requires multiple round-trips to the database. 对于单个查询,我希望
in
大多数情况下都会取胜-仅仅因为创建表然后使用它需要多次往返数据库。
In addition, some databases optimize constant lists (for instance, MySQL does a binary search on values rather than a sequential search). 此外,某些数据库还优化了常量列表(例如,MySQL对值进行二进制搜索,而不是顺序搜索)。
The one thing that will help either version is an index on (col1)
or (col1, col2, col3, col4)
. 可以帮助任一版本的一件事是
(col1)
或(col1, col2, col3, col4)
上的索引。
声明:本站的技术帖子网页,遵循CC BY-SA 4.0协议,如果您需要转载,请注明本站网址或者原文地址。任何问题请咨询:yoyou2525@163.com.