SQL进阶之字符串函数使用及分组的使用SQL函数array_length ,string_to_array ,array_to_string ,array_agg

应用场景
使用group by后需要对数据分组后以分组形式展示,然后还需要对数据去重并获得不重复的值的个数
但是还需要算出有几个,因为统计的是有多少天在下雨,现在查到的是哪一天

使用函数将目标数据变成组展示

array_to_string(array_agg( to_char(detection_time, ‘dd’)), ‘,’) AS rainday : 统计下雨的天数， to_char(detection_time, ‘dd’) 将日期转换为当月的日数， DISTINCT 去重， array_agg 将所有日数聚合为一个数组， array_to_string 将数组转换为逗号分隔的字符串。
这时候查询的数据是有重复的**(统计都有哪几天下雨),我们只需要在to_char前加上DISTINCT** 就可以去除重复

完整SQL

( SELECT COUNT(*) * 5 AS rainFallTime,array_to_string(array_agg(DISTINCT to_char(detection_time, 'dd')), ',') AS rainday,
        SUM(rainfall) AS rainFall,date_trunc('month', detection_time) AS month
        FROM road_environment_info
        WHERE
                observation_point_id = 32
          AND detection_time BETWEEN '2020-04-01' AND '2023-11-01'
          AND rainfall IS NOT NULL
          AND rainfall != 0
        GROUP BY
            date_trunc('month', detection_time)
        ORDER BY
            month
    )

到这拿到就是不重复的所有值,但是还需要算出有几个,因为统计的是有多少天在下雨,现在查到的是哪一天

使用字符串函数和count算出个数

思路:以上面的查询作为子查询,在外层嵌套查询使用array_length(string_to_array(a.rainday, ‘,’), 1) AS rainDayCount计算下雨天数的数量。

具体解释如下：

• string_to_array(a.rainday, ‘,’) : 将逗号分隔的字符串 a.rainday 转换为数组。
• array_length(string_to_array(a.rainday, ‘,’), 1) : 计算数组的长度，其中 1 表示数组的维度。
• AS rainDayCount : 将计算得到的下雨天数数量赋值给别名 rainDayCount 。

完整SQL

SELECT a.rainFallTime,a.rainFall,a.month,array_length(string_to_array(a.rainday, ','), 1) AS rainDayCount
FROM
    ( SELECT COUNT(*) * 5 AS rainFallTime,array_to_string(array_agg(DISTINCT to_char(detection_time, 'dd')), ',') AS rainday,
                SUM(rainfall) AS rainFall,date_trunc('month', detection_time) AS month
        FROM road_environment_info
        WHERE
                observation_point_id = 32
          AND detection_time BETWEEN '2020-04-01' AND '2023-11-01'
          AND rainfall IS NOT NULL
          AND rainfall != 0
        GROUP BY
            date_trunc('month', detection_time)
        ORDER BY
            month
    ) AS a
GROUP BY
    a.rainFallTime, a.rainFall, a.month,a.rainday
ORDER BY
    a.month