优化没有索引的表的删除查询？

Question

Let's say we have the following tables: r(a, b, c) where a is some integer, and the table has 50 rows s(d, e, f, a) sa is a foreign key onto ta each tuple is 400 bytes and s is a multiset since this is a SQL database. s holds about 1 million tuples. when a table is added, the database server automatically creates an index on the primary-‐key columns in the table so for example t has an index on a. However, Oracle does not create indexes on columns that are foreign keys. So s has no index. We want to delete 15 rows from t. So here is what we do: 1) We remove rows from s that have the same a values in t. This takes 10 minutes. 2) Then we run the statement DELETE FROM t WHERE a IN (values we are trying to remove, 15 of them to be exact) This takes about 6 hours.

So my guess is the second step takes a long time because for every tuple in t, we are comparing the value of a to every value within the query list. So how can we optimize this query to get the same result but much faster!?

Answer 1

if you use joins For CRUD Operations will be faster rather than using IN or subquery

For eg:

DELETE FROM t 
INNER JOIN S ON t.a = S.a 

Hope this helps