仅从 PostgreSQL 中 Java Date 类格式的字段按年份分组

Question

我在 PostgreSQL 中工作，我有表名 my_table 架构

CREATE TABLE my_table
(
  my_table_id character varying(50) NOT NULL,
  bar_id character varying(50),
  beer_id character varying(50),
  status_id character varying(50),

  updated_at bigint,

  CONSTRAINT my_table_id PRIMARY KEY (craft_bar_beer_id ),
  CONSTRAINT my_table_bar_id_fk FOREIGN KEY (bar_id)
      REFERENCES table_bar (bar_id) MATCH Unknown
      ON UPDATE NO ACTION ON DELETE NO ACTION,
  CONSTRAINT beer_status_id_fk FOREIGN KEY (status_id)
      REFERENCES status (status_id) MATCH Unknown
      ON UPDATE NO ACTION ON DELETE NO ACTION,
  CONSTRAINT beer_id_fk FOREIGN KEY (beer_id)
      REFERENCES beer (beer_id) MATCH Unknown
      ON UPDATE NO ACTION ON DELETE NO ACTION
);

现在输出是 beer_id 的计数，其状态为“1”和“2”并按日期分组（此处假设提供 2016 年或应显示过去 10 年的记录）

count(beer_id)   status     year
4                  1        2014
5                  1        2015
3                  2        2014
6                  2        2016

但这里updated_at 字段是bigint，它是从Java Date 对象派生的以毫秒为单位的时间：

Date dateValue = null;
SimpleDateFormat formatter = new SimpleDateFormat(dateFormat);
formatter.setTimeZone(TimeZone.getTimeZone("GMT"));
dateValue = formatter.parse(dateStr);
return dateValue.getTime();

我试过的查询：

SELECT count(cbeer.beer_id) AS beers, cbeer.status_id AS Status
FROM my_table cbeer
INNER JOIN bar cb ON cb.bar_id = cbeer.bar_id
WHERE status_id IN ('1','2')
GROUP BY cbeer.status_id;

输出：

beers Status
5      1
1      2

当日期很长并且我希望它按过去 10 年分组时，如何查询以获取日期明智的记录。如果为 ex: 2015 提供年份，则按 2015 分组并获取该年的月份相关记录。

Answer 1

Java Date.getTime() 函数将对象作为 long 返回自 1970 年 1 月 1 日以来的毫秒数。幸运的是，这是 PostgreSQL 使用的同一时期，因此您可以简单地将值除以 1,000 并转换为 timestamp 以拥有对 PostgreSQL 日期和时间函数的完全访问权限：

to_timestamp(updated_at / 1000.)

现在您的查询。过去 10 年：

SELECT count(cbeer.beer_id) AS beers,
       cbeer.status_id AS Status,
       extract('year' from to_timestamp(updated_at / 1000.)) AS year
FROM my_table cbeer
INNER JOIN bar cb ON cb.bar_id = cbeer.bar_id
WHERE status_id IN ('1','2')
  AND extract('year' from to_timestamp(updated_at / 1000.)) > 2005
GROUP BY 2, 3
ORDER BY 2, 3, 1;

当月：

SELECT count(cbeer.beer_id) AS beers,
       cbeer.status_id AS Status,
       extract('month' from to_timestamp(updated_at / 1000.)) AS month
FROM my_table cbeer
INNER JOIN bar cb ON cb.bar_id = cbeer.bar_id
WHERE status_id IN ('1','2')
  AND extract('year' from to_timestamp(updated_at / 1000.)) = 2015
GROUP BY 2, 3
ORDER BY 2, 3, 1;