data comparison and finding optimal solution using SQL - sql

Question

I have 2 tables. One has data for all the guests invited. Other has all the guests who came for the party. How would I check guests that did not come using SQL. Can you make it efficient assuming hypothetically there are 500k guests invited and 300k guests arrived.
If we have to use Script say bash script to automate it. What would the script look like?

Answer 1

Im not really sure that name will be a unique enough key but, here is a query that will get rows from one table that do not exist on another:
SELECT name
FROM tablea
WHERE NOT EXISTS
(SELECT name FROM tableb)
;

Answer 2

Sometimes, the set operators are the simplest:
select name from invited
minus
select name from attended

Answer 3

With a suitable index defined, I'd use an anti-join pattern
SELECT i.guest
FROM invited i
-- anti-join exclude
LEFT
JOIN attended a
ON a.guest = i.guest
WHERE a.guest IS NULL
An equivalent result can be returned with other query patterns.
For example, a NOT EXISTS with a correlated subquery.
SELECT i.guest
FROM invited i
WHERE NOT EXISTS ( SELECT 1
FROM attended a
WHERE a.guest = i.guest
)
The EXPLAIN output will likely show identical query access plans.
For optimum performance of those queries, we'd want an index defined
ON attended (guest)

Answer 4

Based on an earlier comment, I want to demonstrate the difference between a MINUS and the ANTI JOIN
I have a table of INVITED about 1.8M and GUESTS (attended) about 1.1M
Using the MINUS,
select name from invited
minus
select name from attended
---------------------------------------------------------------------------------------
| Id | Operation | Name | Rows | Bytes |TempSpc| Cost (%CPU)| Time |
---------------------------------------------------------------------------------------
| 0 | SELECT STATEMENT | | | | | 14364 (100)| |
| 1 | MINUS | | | | | | |
| 2 | SORT UNIQUE | | 1835K| 10M| 21M| 8977 (1)| 00:00:01 |
| 3 | TABLE ACCESS FULL| INVITED | 1835K| 10M| | 3038 (1)| 00:00:01 |
| 4 | SORT UNIQUE | | 1101K| 6451K| 12M| 5388 (1)| 00:00:01 |
| 5 | TABLE ACCESS FULL| GUESTS | 1101K| 6451K| | 1824 (1)| 00:00:01 |
---------------------------------------------------------------------------------------
Using the ANTI JOIN
SELECT i.name
FROM invited i
-- anti-join exclude
LEFT
JOIN guests g
ON g.name = i.name
WHERE g.name IS NULL
;
----------------------------------------------------------------------------------------
| Id | Operation | Name | Rows | Bytes |TempSpc| Cost (%CPU)| Time |
----------------------------------------------------------------------------------------
| 0 | SELECT STATEMENT | | | | | 7744 (100)| |
|* 1 | HASH JOIN RIGHT ANTI| | 1835K| 21M| 18M| 7744 (1)| 00:00:01 |
| 2 | TABLE ACCESS FULL | GUESTS | 1101K| 6451K| | 1824 (1)| 00:00:01 |
| 3 | TABLE ACCESS FULL | INVITED | 1835K| 10M| | 3038 (1)| 00:00:01 |
----------------------------------------------------------------------------------------
As you can see the ANTI JOIN method does not need to perform the sort and ends up being about half the cost of the MINUS version.