Question

I have to count how many times a number from table2 occurs between the number in range table2.a and table2.b

i.e. we wanna know how many times we have this : a < start < b

I ran the following query :

UPDATE table2
 SET occurrence =  
       (SELECT COUNT(*) FROM table1 WHERE start BETWEEN table2.a AND table2.b);



table2 
ID   a   b    occurrence
1    1   10
2    1   20
3    1   25
4    2   30


table1 
ID start col1 col2 col3
1   1
2   7
3  10
4  21
5  25
6  27
7  30

table2 as

• 3 indexes on a, b and occurrence
• 1567 rows (so we will SELECT COUNT(*) over table2 1567 times..)
• ID column as PK

table1 as

• 1 index on start
• 42,000,000 rows
• Column start was “ordered by column start”
• ID column as PK

==> it took 2.5hours to do 2/3 of it. I need to speed this up… any suggestions ? 🙂

Answer 1

You could try to add the id column to the index on table 1:

CREATE INDEX start_index ON table1 (start,id);

And rewrite the query to

UPDATE table2
 SET occurrence =  
       (SELECT COUNT(id) FROM table1 WHERE start BETWEEN table2.a AND table2.b);

This is called “covering index”: http://www.simple-talk.com/sql/learn-sql-server/using-covering-indexes-to-improve-query-performance/

-> The whole query on table 1 can be served through the data in the index -> no additional page lookup for the actual record.