在PostgreSQL中使用公共表表达式（CTE）构建表时，如何获取唯一字段？

我试图生成一系列时间戳，使用PostGreSQL计算每个时间戳的星期几，并只选择星期一到星期六的时间戳

以下是我迄今为止的尝试：

WITH
candidates AS (SELECT * FROM generate_series(
  date_trunc('hour', CURRENT_TIMESTAMP - INTERVAL '2 hours'),
  date_trunc('hour', CURRENT_TIMESTAMP - INTERVAL '2 hours')+INTERVAL '1 week',
  INTERVAL '1 hour')),
candidows AS (SELECT EXTRACT(DOW FROM generate_series) FROM candidates),
candidatable AS (SELECT b.date_part AS dow, a.generate_series as start_booking FROM candidates a, candidows b)
SELECT * FROM candidatable WHERE dow != 0;

返回的视图包含一个行为与我期望的一样的dow列，但start_booking列只包含相同的值：

 dow |     start_booking
-----+------------------------
   2 | 2017-09-19 11:00:00+01
   2 | 2017-09-19 11:00:00+01
   2 | 2017-09-19 11:00:00+01
   2 | 2017-09-19 11:00:00+01
   2 | 2017-09-19 11:00:00+01
   2 | 2017-09-19 11:00:00+01
   2 | 2017-09-19 11:00:00+01
   2 | 2017-09-19 11:00:00+01
   2 | 2017-09-19 11:00:00+01
   2 | 2017-09-19 11:00:00+01
   2 | 2017-09-19 11:00:00+01
   2 | 2017-09-19 11:00:00+01
   2 | 2017-09-19 11:00:00+01
   3 | 2017-09-19 11:00:00+01
   3 | 2017-09-19 11:00:00+01
   3 | 2017-09-19 11:00:00+01

如果我将最后一条SELECT语句替换为：

SELECT * FROM candidates;

然后，正如我所预期的那样，我试图在上面查看的数据是连续的：

    generate_series
------------------------
 2017-09-19 11:00:00+01
 2017-09-19 12:00:00+01
 2017-09-19 13:00:00+01
 2017-09-19 14:00:00+01
 2017-09-19 15:00:00+01
 2017-09-19 16:00:00+01

---

如何从候选人处获取数据。生成系列以候选表格的形式显示。开始预订？

猜测-您是否在尝试：

WITH
candidates AS (SELECT generate_series start_booking, EXTRACT(DOW FROM generate_series) dow FROM generate_series(
  date_trunc('hour', CURRENT_TIMESTAMP - INTERVAL '2 hours'),
  date_trunc('hour', CURRENT_TIMESTAMP - INTERVAL '2 hours')+INTERVAL '1 week',
  INTERVAL '1 hour'))
SELECT * FROM candidates WHERE dow != 0;

---

好吧，这是一个更好的行为，但我不希望结果包含一个名为“generate_series”的列。我可以重命名它吗？eddited包含通缉犯别名