mysql - Mysql优化避免表扫描

Question

请考虑下表：

_____________________
|   sentence_word   |
|---------|---------|
| sent_id | word_id |
|---------|---------|
| 1       | 1       |
| 1       | 2       |
| ...     | ...     |
| 2       | 4       |
| 2       | 1       |
| ...     | ...     |

有了这个表结构，我想存储句子的单词。现在我想找出句子中哪些单词与特定单词一起出现。结果应如下所示：

_____________________
| word_id | counted |
|---------|---------|
| 5       | 1000    |
| 7       | 800     |
| 3       | 600     |
| 1       | 400     |
| 2       | 100     |
| ...     | ...     |

查询如下所示：

SELECT
    word_id,
    COUNT(*) AS counted
FROM sentence_word
WHERE sentence_word.sent_id IN (SELECT
    sent_id
    FROM sentence_word
    WHERE word_id = [desired word]
)
AND word_id != [desired word]
GROUP BY word_id
ORDER BY counted DESC;

查询正常工作，但它始终扫描整个表。我为 sent_id 和 word_id 创建了一个索引。你有什么想法可以优化它，它不需要一直扫描整个表吗？

Answer 1

你可以尝试这样的自我加入：

SELECT COUNT(DISTINCT sw1.word_id)
FROM sentence_word sw1
JOIN sentence_word sw2 ON (
    sw1.sent_id = sw2.sent_id
    AND sw2.word_id = [your word id]
)
WHERE sw1.word_id != [your word id]

或者甚至更好

SELECT COUNT(DISTINCT sw1.word_id)
FROM sentence_word sw1
JOIN sentence_word sw2 ON (
    sw1.sent_id = sw2.sent_id
    AND sw2.word_id = [your word id]
    AND sw2.word_id != sw1.word_id
)