postgresql联盟

我在postgresql中有这些数据

  id      | rate      | hrv | activity | orientation |       timestamp        | user_id | status 
----------+-----------+-----+----------+-------------+------------------------+---------+--------------
 66764728 |        72 |   1 |        0 |          90 | 2010-06-10 18:54:54+00 |     397 | t
 66764729 |        72 |   1 |        0 |          90 | 2010-06-10 18:55:09+00 |     397 | t
 66764730 |        76 |   1 |        0 |          90 | 2010-06-10 18:55:23+00 |     397 | t
 66766058 |        68 |   1 |        0 |          90 | 2010-06-10 18:55:38+00 |     397 | t
 66766059 |        72 |   1 |        0 |          90 | 2010-06-10 18:55:53+00 |     397 | t
 66766063 |        80 |   1 |        0 |           0 | 2010-06-10 18:56:51+00 |     397 | t
 66766085 |       100 |   1 |        0 |           0 | 2010-06-10 18:57:06+00 |     397 | t
 66766091 |        -1 |  -1 |     1558 |          90 | 2010-06-10 18:58:34+00 |     397 | t
 66766118 |        -1 |  -1 |     2055 |           0 | 2010-06-10 18:58:49+00 |     397 | t
 66766119 |        -1 |  -1 |     2869 |          90 | 2010-06-10 18:59:03+00 |     397 | t
 66766121 |        -1 |  -1 |     3187 |          90 | 2010-06-10 18:59:18+00 |     397 | t
 66766120 |        -1 |  -1 |     3302 |           0 | 2010-06-10 18:59:33+00 |     397 | t
 66766122 |        -1 |  -1 |     2222 |           0 | 2010-06-10 18:59:47+00 |     397 | t
 66766133 |        60 |   7 |        0 |           0 | 2010-06-10 19:00:16+00 |     397 | t
 66766134 |        64 |   1 |        0 |           0 | 2010-06-10 19:00:31+00 |     397 | t
 66766135 |        72 |   1 |        0 |           0 | 2010-06-10 19:00:46+00 |     397 | t
 66766137 |        72 |   0 |        0 |           0 | 2010-06-10 19:01:15+00 |     397 | t
 66766155 |       132 |   1 |        0 |           0 | 2010-06-10 19:01:59+00 |     397 | t
 66766159 |        -1 |  -1 |     1858 |          90 | 2010-06-10 19:02:58+00 |     397 | t

我该怎么去

( ROUND(AVG(rate),1) AS avg_rate, hrv WHERE rate <> -1 ) UNION ( ROUND(AVG(activity),1) AS avg_activity, hrv WHERE activity <> -1 )

---

如果我正确理解您的需求，这应该可以做到：

select round(avg(case
                     when rate <> -1 then rate
                     else null
                 end), 1) as avg_rate
     , round(avg(case
                     when activity <> -1 then activity
                     else null
                 end), 1) as avg_activity
     , round(avg(hrv)) as avg_hrv
  from my_table

---

你错过了WHERE条件。每个查询都不同。AVGrate有一组行，AVGactivity有其他行。它们不是同一组。否则，我会很好地进行直接查询。除此之外，我希望这些列显示在一行中，但显示为两列。@Ramonrails-我相信我所做的应该没有经过测试，因此可能有一些语法错误，但希望没有：。这对我来说很有效。但是我的SQL更大。这是我问的一个非常简化的版本谢谢。这称为内联视图，以备将来参考。如何获取avgrate、avgactivity和avghrv，其中所有三个结果集都不同于它们导出平均值的结果集？case语句允许您通过将不希望考虑的行计数为null，对不同部分进行平均值计算，这并没有计入平均值。这里，avg_rate是表中所有行的平均值，例如，其中rate不是-1。是的，我知道CASE语句。不过我问了一些不同的问题。没关系，我现在明白了。谢谢陆克文插话：哦，当然。对不起，我帮不上忙了！

SELECT a.avg_rate
     , b.avg_activty
  FROM
(SELECT ROUND(AVG(rate),1) AS avg_rate
 WHERE rate <> -1) a,
(SELECT ROUND(AVG(activity),1) AS avg_activity
 WHERE activity <> -1) b;

select round(avg(case
                     when rate <> -1 then rate
                     else null
                 end), 1) as avg_rate
     , round(avg(case
                     when activity <> -1 then activity
                     else null
                 end), 1) as avg_activity
     , round(avg(hrv)) as avg_hrv
  from my_table