SELECT and ORDER with several indexes on a table

Question

Hi,
I am looking for the best way to select rows from a table if they satisfy a condition depending on several indexes, or from a join.
I am currently doing it with unions, like that:
SELECT article_id, article_date, score(1) score
FROM articles
WHERE CONTAINS(article_body,'searchQuery',1) > 0
UNION
SELECT article_id, article_date, score(1) score
FROM articles
WHERE CONTAINS(article_title,'searchQuery',1) > 0
UNION
SELECT article_id, article_date, 100
FROM articles
WHERE article_id IN (SELECT article_id FROM article_keywords WHERE keyword = 'searchQuery')
ORDER BY score DESC, article_date DESC
But I was wondering whether it would be better performance-wise to do it with conditional expressions like
SELECT article_id, article_date, score(1) scoreBody, score(2) scoreTitle
FROM articles
WHERE
CONTAINS (article_body,'searchQuery',1) > 0
OR CONTAINS (article_title,'searchQuery',2) > 0
OR article_id IN (SELECT article_id FROM article_keywords WHERE keyword = 'searchQuery')
ORDER BY scoreHeadline DESC, scoreBody DESC, article_date DESC;
Also, if doing it with conditional expressions, I don't see how to avoir having to declare several "scores".
In a more general way, I don't see exactly how to consider the results coming from the join with the article_keywords table in terms of score (right now I'm giving them 100, so they're coming up first).
I am completely new to indexes, so my apologies if anything in this SQL is hurting just looking at it.
Thanks for reading, and for any help you can bring.

Answer 1

You can check out the explain plan for the two queries. I did a small test for employees table . The first query scans the table 3 times, while the second does only once.
hr@XE> EXPLAIN PLAN FOR
  2  SELECT * FROM employees
  3  WHERE last_name like 'Joh%'
  4  UNION
  5  SELECT * FROM employees
  6  WHERE last_name like 'Pet%'
  7  UNION
  8  SELECT * FROM employees
  9  WHERE last_name like 'King%';
Explained.
Elapsed: 00:00:03.06
hr@XE> select * from table(dbms_xplan.display);
PLAN_TABLE_OUTPUT
Plan hash value: 1534505908
| Id  | Operation                     | Name        | Rows  | Bytes | Cost (%CPU)| Time     |
|   0 | SELECT STATEMENT              |             |     3 |   204 |     9  (78)| 00:00:01 |
|   1 |  SORT UNIQUE                  |             |     3 |   204 |     9  (78)| 00:00:01 |
|   2 |   UNION-ALL                   |             |       |       |            |          |
|   3 |    TABLE ACCESS BY INDEX ROWID| EMPLOYEES   |     1 |    68 |     2   (0)| 00:00:01 |
|*  4 |     INDEX RANGE SCAN          | EMP_NAME_IX |     1 |       |     1   (0)| 00:00:01 |
|   5 |    TABLE ACCESS BY INDEX ROWID| EMPLOYEES   |     1 |    68 |     2   (0)| 00:00:01 |
|*  6 |     INDEX RANGE SCAN          | EMP_NAME_IX |     1 |       |     1   (0)| 00:00:01 |
|   7 |    TABLE ACCESS BY INDEX ROWID| EMPLOYEES   |     1 |    68 |     2   (0)| 00:00:01 |
|*  8 |     INDEX RANGE SCAN          | EMP_NAME_IX |     1 |       |     1   (0)| 00:00:01 |
Predicate Information (identified by operation id):
   4 - access("LAST_NAME" LIKE 'Joh%')
       filter("LAST_NAME" LIKE 'Joh%')
   6 - access("LAST_NAME" LIKE 'Pet%')
       filter("LAST_NAME" LIKE 'Pet%')
   8 - access("LAST_NAME" LIKE 'King%')
       filter("LAST_NAME" LIKE 'King%')
hr@XE> explain plan for
  2  select * from employees
  3  where last_name like 'Joh%' or last_name like 'Pet%' or last_name like 'King%';
Explained.
Elapsed: 00:00:00.01
hr@XE> select * from table(dbms_xplan.display);
PLAN_TABLE_OUTPUT
Plan hash value: 1445457117
| Id  | Operation         | Name      | Rows  | Bytes | Cost (%CPU)| Time     |
|   0 | SELECT STATEMENT  |           |     3 |   204 |     3   (0)| 00:00:01 |
|*  1 |  TABLE ACCESS FULL| EMPLOYEES |     3 |   204 |     3   (0)| 00:00:01 |
Predicate Information (identified by operation id):
   1 - filter("LAST_NAME" LIKE 'Joh%' OR "LAST_NAME" LIKE 'Pet%' OR
              "LAST_NAME" LIKE 'King%')