我正在尝试创建一个具有30天记录历史记录的选择,其中values列得到计数,我只是希望它可以添加每天返回的值。
查询
SELECT
date_update AS Time,
SUM(COUNT(values)) as "Value"
FROM
tb_get_metrics
WHERE
data_update >= CURRENT_DATE - 30
GROUP BY Time
ORDER BY Time
示例输出:
Time Value
2019-10-14 09:46:54.789772 30
2019-10-15 09:46:54.789772 50
2019-10-16 09:46:54.789772 70
选择*来自tb_get_metrics
date_update(TimeStamp) value(String)
2019-10-14 09:46:54.789772 apple
2019-10-14 09:46:55.789772 apple
2019-10-14 09:46:56.789772 apple
2019-10-14 09:46:57.789772 apple
2019-10-14 09:46:58.789772 apple
2019-10-14 09:46:59.789772 apple
2019-10-14 09:47:00.789772 apple
2019-10-14 09:46:01.789772 apple
2019-10-14 09:46:02.789772 apple
2019-10-14 09:46:03.789772 apple
2019-10-14 09:46:04.789772 apple
2019-10-14 09:46:05.789772 apple
2019-10-15 09:46:03.789772 potato
2019-10-15 09:46:04.789772 potato
2019-10-15 09:46:05.789772 potato
...
答案 0 :(得分:1)
您需要舍入日期的时间:
SELECT
date_update::date AS "Date",
COUNT(values) as "Value"
FROM
tb_get_metrics
WHERE
data_update >= CURRENT_DATE - 30
GROUP BY Time
ORDER BY Time
答案 1 :(得分:1)
仅从每个时间戳中获取日期部分以进行分组:
SELECT
date_update::date AS Time,
COUNT(values) AS "Value"
FROM tb_get_metrics
WHERE date_update >= CURRENT_DATE - 30
GROUP BY Time
ORDER BY Time
答案 2 :(得分:1)
通过截断时间戳记的时间,您可以仅关注日期,这是您的WHERE
子句试图做到的:
postgres=# SELECT
date_update::date AS "Date",
COUNT(values) as "Count"
FROM
tb_get_metrics
WHERE
date_update::date >= CURRENT_DATE - 30
GROUP BY "Date"
ORDER BY "Date";
Date | Count
------------+-------
2019-10-14 | 24
2019-10-15 | 6
(2 rows)