rinri
/
databases-course


			
				
					
						
						
							
							# Before indexing

EXPLAIN ANALYZE SELECT * from customer WHERE name='John Smith';
                                              QUERY PLAN
------------------------------------------------------------------------------------------------------
 Seq Scan on customer  (cost=0.00..429.00 rows=3 width=211) (actual time=0.299..3.124 rows=3 loops=1)
   Filter: (name = 'John Smith'::text)
   Rows Removed by Filter: 9997
 Planning Time: 0.129 ms
 Execution Time: 3.146 ms
(5 rows)

# After indexing using btree(name):

CREATE INDEX idx_name on customer using btree(name);
EXPLAIN ANALYZE SELECT * from customer WHERE name='John Smith';
                                                   QUERY PLAN

-------------------------------------------------------------------------------------------------------
----------
 Bitmap Heap Scan on customer  (cost=4.31..15.45 rows=3 width=211) (actual time=0.051..0.056 rows=3 loo
ps=1)
   Recheck Cond: (name = 'John Smith'::text)
   Heap Blocks: exact=3
   ->  Bitmap Index Scan on idx_name  (cost=0.00..4.31 rows=3 width=0) (actual time=0.044..0.044 rows=3
 loops=1)
         Index Cond: (name = 'John Smith'::text)
 Planning Time: 0.423 ms
 Execution Time: 0.086 ms

# A complicated query with like doesn't benefit from hash indexing

EXPLAIN ANALYZE SELECT * FROM customer WHERE address LIKE '%Warren%Port%';
                                              QUERY PLAN
------------------------------------------------------------------------------------------------------
 Seq Scan on customer  (cost=0.00..429.00 rows=1 width=211) (actual time=3.544..4.110 rows=1 loops=1)
   Filter: (address ~~ '%Warren%Port%'::text)
   Rows Removed by Filter: 9999
 Planning Time: 0.125 ms
 Execution Time: 4.131 ms
(5 rows)

The query can be improved and benefit from the btree indexing:
EXPLAIN ANALYZE SELECT * FROM customer WHERE name >= 'Lisa' and name <= 'Lisb';
                                                    QUERY PLAN

-------------------------------------------------------------------------------------------------------
------------
 Bitmap Heap Scan on customer  (cost=4.41..45.44 rows=12 width=211) (actual time=0.068..0.202 rows=90 l
oops=1)
   Recheck Cond: ((name >= 'Lisa'::text) AND (name <= 'Lisb'::text))
   Heap Blocks: exact=83
   ->  Bitmap Index Scan on idx_name  (cost=0.00..4.41 rows=12 width=0) (actual time=0.042..0.042 rows=
90 loops=1)
         Index Cond: ((name >= 'Lisa'::text) AND (name <= 'Lisb'::text))
 Planning Time: 0.174 ms
 Execution Time: 0.233 ms

Before indexing address:

EXPLAIN ANALYZE SELECT * FROM customer WHERE address='USS Oneill FPO AE 11865';

                                              QUERY PLAN
------------------------------------------------------------------------------------------------------
 Seq Scan on customer  (cost=0.00..429.00 rows=1 width=211) (actual time=2.639..2.639 rows=0 loops=1)
   Filter: (address = 'USS Oneill FPO AE 11865'::text)
   Rows Removed by Filter: 10000
 Planning Time: 0.084 ms
 Execution Time: 2.657 ms
(5 rows)

After indexing address:
CREATE INDEX idx_address on customer using HASH(address);
EXPLAIN ANALYZE SELECT * FROM customer WHERE address='USS Oneill FPO AE 11865';
                                                       QUERY PLAN

-------------------------------------------------------------------------------------------------------
-----------------
 Index Scan using idx_address on customer  (cost=0.00..8.02 rows=1 width=211) (actual time=0.016..0.016
 rows=0 loops=1)
   Index Cond: (address = 'USS Oneill FPO AE 11865'::text)
 Planning Time: 0.224 ms
 Execution Time: 0.035 ms
(4 rows)