Question

完全披露：我已经看到mySQL的这个问题有1个变体，而PostgreSQL的答案并不能让我满意。

我有2张桌子：评论＆amp;商家。在评论表中，针对此问题的唯一3个相关列是＆＃39; business_id＆＃39;，＆＃39; date＆＃39; （yyyy-mm-dd）和星号（1-5），主键是（review_id）。在商家信息表格中，相关列是＆＃39; business_id＆＃39;，＆＃39; year＆＃39;和＆＃39; month＆＃39;＆＃39; ＆＃39;年＆＃39;和＆＃39;月＆＃39;因为商家信息系统中还有另一列名为＆＃39; review_count＆＃39;，它代表了每年每个月收到的商家评论数量。因此，此表的复合主键是（business_id，year，month）。

基本上，我正在尝试在商家信息表中创建一个列，其中每个月的每个月都会收到一个商家的平均评分（以星号表示）。

以下查询为我提供了我想要的确切结果：

SELECT round(CAST(AVG(stars) AS NUMERIC), 2)
                                    FROM reviews_for_trending_businesses
                                    WHERE business_id IN (SELECT DISTINCT(business_id)
                                                                          FROM trending_businesses_v2)
                                    GROUP BY business_id, EXTRACT("year" FROM reviews_for_trending_businesses.date), EXTRACT('month' FROM reviews_for_trending_businesses.date);

此代码返回列以及我要插入业务表的所有正确值。

但是，当我尝试实际更新表时，我收到一条错误，指出用作表达式的子查询返回了多行。这是我尝试更新的代码：

UPDATE trending_businesses_v2
SET avg_monthly_rating = (SELECT round(CAST(AVG(stars) AS NUMERIC), 2)
                                        FROM reviews_for_trending_businesses
                                        WHERE business_id IN (SELECT DISTINCT(business_id)
                                                                              FROM trending_businesses_v2)
                                        GROUP BY business_id, EXTRACT("year" FROM reviews_for_trending_businesses.date), EXTRACT('month' FROM reviews_for_trending_businesses.date);

我也尝试过其他一些解决方案，包括使用连接，但仍会遇到类似的错误。

更新：仍然没有答案，但越来越接近：仍然无法弄清楚我在哪里出错了。我也不明白为什么我必须分组＆＃39; rtb.date＆＃39;在这里，如果我只从中提取值（如果我没有，则返回错误）。

UPDATE trending_businesses_v2 tb
    SET avg_monthly_rating = t.val
    FROM (SELECT business_id, EXTRACT("year" FROM rtb.date) AS year, EXTRACT('month' FROM rtb.date) AS month, round(CAST(AVG(stars) AS NUMERIC), 2) as val
          FROM reviews_for_trending_businesses rtb
          WHERE business_id IN (SELECT DISTINCT(business_id)

                                FROM trending_businesses_v2
                               )
          GROUP BY business_id, year, month, rtb.date
         ) t
    WHERE t.business_id = tb.business_id AND
          t.year = tb.year AND t.month = tb.month;

Answer 1

您需要匹配行，可能是使用业务ID和日期。像这样：

UPDATE trending_businesses_v2 tb
    SET avg_monthly_rating = t.val
    FROM (SELECT business_id, date_trunc('month', rtb.date) as yyyymm, round(CAST(AVG(stars) AS NUMERIC), 2) as val
          FROM reviews_for_trending_businesses rtb
          WHERE business_id IN (SELECT DISTINCT(business_id)

                                FROM trending_businesses_v2
                               )
          GROUP BY business_id, date_trunc('month', rtb.date)
         ) t
    WHERE t.business_id = tb.business_id AND
          t.yyyymm = tb.?;

如何设置新列等于子查询PostGreSQL

1 个答案: