Reputation: 38213
I have 2 tables:
tblValidItems - | - tblItems
validID itemID
------- ------
3 1
5 2
6 3
... 4
~ 8 K items 5
.....
~ 20 K items
My query is to select certain things in tblItems
that are also in tblValidItems
:
SELECT tblItems.itemID FROM tblItems
JOIN tblValidItems ON tblItems.itemID = tblValidItems.validID
I tried the query with and without an index on both tables, but the results varied vary little:
This surprised me because I thought an index would dramatically effect the speed of this query. Why dons't it?
Upvotes: 4
Views: 452
Reputation: 1269503
I am guessing that the query is dominated by the time to return the 8,000 values and not the time to find the rows.
Indexes are most useful when you are reducing the size of the data you are working with. The reduction from 20k rows to 8k is not particularly signficant.
Upvotes: 5
Reputation: 22241
Sqlite primary keys are indexed by default. You are joining on an indexed primary key anyway.
Whenever you have doubts on how sqlite will work with your query, use EXPLAIN QUERY PLAN
Upvotes: 4