相关疑难解决方法(0)

在 PG v9.4 及更高版本上，我想将一些数据（基于 SELECT 语句）导出为JSON 数组。

标准 json_agg 返回我想要的内容，但它返回对象数组（其中对象键是列名称）例如此查询：

SELECT json_agg(data_rows)
FROM (
        -- in memory table sample
    SELECT * FROM 
    (VALUES
        ('John',now(),1940,'Winston','Lennon'),
        ('Paul',now(),1942,'','McCartney'),
        ('George',now(),1943,NULL,'Harrison'),
        ('Ringo',now(),1940,'my passions are ring,drum and shades','Starr')
    ) AS x("FirstName", "CurrentDt", "BirthYear", "MiddleName", "LastName")
        ORDER BY "BirthYear" DESC, "FirstName" DESC
) AS data_rows

返回以下内容：

[
 {"FirstName":"George","CurrentDt":"2016-09-12T13:13:07.862485-04:00","BirthYear":1943,"MiddleName":null,"LastName":"Harrison"}, 
 {"FirstName":"Paul","CurrentDt":"2016-09-12T13:13:07.862485-04:00","BirthYear":1942,"MiddleName":"","LastName":"McCartney"}, 
 {"FirstName":"Ringo","CurrentDt":"2016-09-12T13:13:07.862485-04:00","BirthYear":1940,"MiddleName":"my passions are ring,drum and shades","LastName":"Starr"}, 
 {"FirstName":"John","CurrentDt":"2016-09-12T13:13:07.862485-04:00","BirthYear":1940,"MiddleName":"Winston","LastName":"Lennon"}
]

但我想要的是：

[
 ["George","2016-09-12T13:13:07.862485-04:00",1943,null,"Harrison"}, 
 ["Paul","2016-09-12T13:13:07.862485-04:00",1942,"","McCartney"}, 
 ["Ringo","2016-09-12T13:13:07.862485-04:00",1940,"my passions are ring,drum and shades","Starr"}, 
 ["John","2016-09-12T13:13:07.862485-04:00",1940,"Winston","Lennon"}
]

我尝试使用这里提到的技巧首先将行转换为 hstore，但这样做的问题是没有保留列排序......所以这个查询：

SELECT json_agg(avals(hstore(data_rows))) …

我正在将一个庞大的表按计数汇总到一个新表中，在该表中我想将所有空字符串更改为NULL，并同时对一些列进行类型转换。我通读了一些帖子，但找不到查询，这使我可以在单个查询中跨所有列进行查询，而无需使用多个语句。

让我知道是否可以遍历所有列，并将单元格替换为具有null的空字符串。

参考：如何使用SQL Server将空格转换为空值？