Bagaimana cara mendapatkan perintah postgres 'nth_value' yang setara di pyspark Hive SQL?

Opsi alternatif adalah row_number() dan fungsi jendela bersyarat:

select
    name,
    weight,
    coalesce(
        max(case when rn = 4 then weight end) over(order by rn),
        99.9
    ) imagined_weight
from (select c.*, row_number() over(order by weight) rn from cats c) c