我有以下sql查询,它根据topicName列进行gruping(也进行了一些除法操作)。 我想为每个分组主题获得2行而不是全部。
SELECT wwt.topicName, t.topic_cnt as sumOfWordsInTopic,
wwt.word, wwt.wordCount,
(wwt.wordCount / t.topic_cnt) AS wordProbability
FROM weightallofwordsintopic as wwt JOIN
(SELECT topicName, sum(wordCount) AS topic_cnt
FROM weightallofwordsintopic
GROUP BY topicName
) t
ON wwt.topicName = t.topicName
weightallofwordsintopic table is as;
topicName | word | wordCount
---
topic0 | word1 | 10
topic0 | word2 | 20
topic0 | word3 | 30
topic0 | word4 | 40
topic0 | word5 | 50
topic0 | word6 | 60
topic1 | word7 | 10
topic1 | word8 | 20
topic1 | word9 | 30
topic1 | word10 | 40
topic1 | word11 | 50
topic1 | word12 | 60
topic2 | word13 | 10
topic2 | word14 | 20
topic2 | word15 | 30
topic2 | word16 | 40
topic2 | word17 | 50
topic2 | word18 | 60
我想要输出(根据它们的重量排序,但在这里我只是放一个样本(选择上面的查询返回一些不同的列)) 我想根据列中的权重将上面的查询限制为每个分组的topicName的2行。
topicName | word | wordCount
topic0 | 1 | 60
topic0 | 1 | 50
topic1 | 1 | 60
topic1 | 1 | 50
topic2 | 1 | 60
topic2 | 2 | 50
答案 0 :(得分:0)
在MySQL中,最简单的方法可能是使用变量:
SELECT t.*
FROM (SELECT wwt.topicName, t.topic_cnt as sumOfWordsInTopic, wwt.word, wwt.wordCount,
(wwt.wordCount / t.topic_cnt) AS wordProbability,
(@rn := if(@t = wwt.topicName, @rn + 1,
if(@t := wwt.topicName, 1, 1)
)
) as rn
FROM weightallofwordsintopic as wwt JOIN
(SELECT topicName, sum(wordCount) AS topic_cnt
FROM weightallofwordsintopic
GROUP BY topicName
) t
ON wwt.topicName = t.topicName CROSS JOIN
(SELECT @t := '', @rn := 0) params
ORDER BY wwt.topicName, wwt.wordCount DESC
) t
WHERE rn <= 2;
答案 1 :(得分:0)
我自己很新,但我相信如果你使用TOP,那么你的SELECT语句就可以了。就像是: &#39; SELECT TOP 2 wwt.topicName等。