Spark SQL：datediff 不会显示结果

Question

无法让Spark SQL datediff工作，也就是说，它运行但不会显示结果。

我尝试了各种方法，最近一次是嵌套

select 
    guid,
    first, 
    last,
    datediff(last_string, first_string)
from (
    select 
        guid,
        first,
        last,
        cast(first as string) as first_string,
        cast(last as string) as last_string
    from (
        select 
            guid, 
            min(entry_date) as first, 
            max(entry_date) as last
        from my_table
        group by guid
    )
)

此查询有效(显示所有 4 列(，但添加日期差异，它只显示 3：guid、第一个、最后一个

select 
    guid,
    first,
    last,
    cast(first as string) as first_string,
    cast(last as string) as last_string
from (
    select 
        guid, 
        min(entry_date) as first, 
        max(entry_date) as last
    from my_table
    group by guid
)

编辑：我已经验证了日期差异有效

select datediff('2012-12-31', '2011-12-31')
=> 366

Answer 1

尝试为括号提供别名 - 最高的选择可能无法正确"看到"您期望的列。