Question

我正在尝试提出一个查询，该查询计算按年分组的行，我将从TIMESTAMP和airline_id获取。到目前为止，我已经想出了这个：

SELECT airline_id, extract(year from departure)::integer as year,
(SELECT count(*) FROM tickets AS t2 LEFT JOIN flights AS f2 ON t2.flight_id = f2.flight_id WHERE extract(year from f2.departure) = year AND f2.airline_id = f.airline_id)
FROM tickets AS t
LEFT JOIN flights AS f ON t.flight_id = f.flight_id
GROUP BY airline_id, year
ORDER BY airline_id ASC, year ASC;

似乎它会正常工作，除非它抛出此错误：

错误：列“年”不存在

我尝试将其引用为：

SELECT airline_id, extract(year from departure)::integer as year,
(SELECT count(*) FROM tickets AS t2 LEFT JOIN flights AS f2 ON t2.flight_id = f2.flight_id WHERE extract(year from f2.departure) = t.year AND f2.airline_id = f.airline_id)
FROM tickets AS t
LEFT JOIN flights AS f ON t.flight_id = f.flight_id
GROUP BY airline_id, year
ORDER BY airline_id ASC, year ASC;

但它做同样的事情。您也无法再次调用该函数，因为

错误：子查询使用来自外部查询的未分组列“f.departure”

那么我该如何从子查询中引用“year”列？

Answer 1

让我们从这个问题开始：

SELECT f.airline_id,  -- I speculate that the airline comes from the flights table
       extract(year from f.departure) as year,
       COUNT(*)
FROM tickets t LEFT JOIN
     flights f 
     ON t.flight_id = f.flight_id
GROUP BY f.airline_id, year
ORDER BY airline_id ASC, year ASC;

它能做到你想要的吗？您的子查询似乎没有任何用处。

如何从子查询引用函数结果列？

1 个答案: