max SQL (hive) 中的嵌套 avg 命令答案

【问题标题】：nested avg command in max SQL (hive)max SQL (hive) 中的嵌套 avg 命令
【发布时间】：2018-03-07 08:43:13
【问题描述】：

这是我的桌子（间距只是为了更好地查看）

month   country   temp
1       AU        15
1       AU        20
2       AU        10
3       AU        20
3       AU        20
3       AU        15

1       CZ        10
1       CZ         5
1       CZ         3
2       CZ        15
2       CZ        15
3       CZ        20
3       CZ        10

1       DE         8
1       DE         2
2       DE        16
2       DE        12
3       DE        21
3       DE        19

我需要显示每月最高平均温度，并使用 SQL 在 HIVE 中显示国家/地区。像这样输出

month   country   max_avg_temp
1       AU        17.5
2       CZ        15
3       DE        20

所以我需要按国家和月份计算 avg(temp) 组，然后找到每个月的最大值。

我试过这样的：

SELECT month, country, MAX(avg_temp) AS max_avg_temp
FROM (SELECT month, country, AVG(temp) as avg_temp FROM temperature
GROUP BY month, country) alias
GROUP BY month;

我也想过这样的事情：

SELECT *
, RANK() OVER (PARTITION BY month ORDER BY avg_temp DESC) AS rank
FROM (SELECT month, country, AVG(temp) AS avg_temp FROM temperature
GROUP BY month, country)
ORDER BY rank
LIMIT 12;

但它不起作用（一些错误）。平均值的嵌套选择工作正常，但我不知道如何显示最大值。

提前致谢

【问题讨论】：

您可以将您的最后一组编辑为GROUP BY month, country 吗？
@HarunKARATAŞ 你的意思是这样的：SELECT month, country, MAX(avg_temp) AS max_avg_temp FROM (SELECT month, country, AVG(temp) as avg_temp FROM temperature GROUP BY month, country) alias GROUP BY month, country; 如果是这样，它只是按月和国家订购所有东西，没有找到每月最大数量

标签： sql hive nested max average

【解决方案1】：

使用row_number() 过滤平均值最高的记录

SELECT month, country, avg_temp
  FROM (  SELECT month,
                 country,
                 AVG (temp) AS avg_temp,
                 ROW_NUMBER ()
                    OVER (PARTITION BY month ORDER BY AVG (temp) DESC)
                    rnk
            FROM temperature
        GROUP BY month, country)  t
 WHERE rnk = 1;

【讨论】：

【解决方案2】：

只是愚蠢的错误，我忘记了一些“别名” - AS t

SELECT *
, RANK() OVER (PARTITION BY mesic ORDER BY avg_temp DESC) AS rank
FROM (SELECT mesic, stat, AVG(teplota) AS avg_temp FROM teploty
GROUP BY mesic, stat) AS t
ORDER BY rank, mesic
limit 12;

它确实显示了每个飞蛾的最高平均温度

+--------------+-------------+---------------------+-------+--+
| t.month      | t.country   |   t.avg_temp        | rank  |
+--------------+-------------+---------------------+-------+--+
| 1            | AS          | 28.1482513153823    | 1     |
| 2            | AS          | 28.268547273982037  | 1     |
| 3            | AS          | 28.305594514570675  | 1     |
| 4            | PW          | 28.14697855750485   | 1     |
| 5            | PW          | 28.15864931145062   | 1     |
| 6            | MH          | 28.085874246339472  | 1     |
| 7            | AZ          | 29.07906528274055   | 1     |
| 8            | TX          | 28.335826966364785  | 1     |
| 9            | MH          | 28.220327304048165  | 1     |
| 10           | MH          | 28.22359756605796   | 1     |
| 11           | MH          | 28.125538329026583  | 1     |
| 12           | AS          | 27.996042413381048  | 1     |
+--------------+-------------+---------------------+-------+--+

【讨论】：