PostgresQL：在具有现有连接的 Unnest 上使用 Where 子句

Question

我已经努力解决了此查询的一些数据格式问题，但无法全神贯注地思考如何限制 unnest 中的月份列以排除未来日期。来源 table 包含不可靠的未来数据，因此我想将其排除。

select fbr.organization, fbr.year, a.accounttype, unnest(array[concat(year,'0101'), concat(year,'0201'), concat(year,'0301'), concat(year,'0401'), concat(year,'0501'), concat(year,'0601'), concat(year,'0701'), concat(year,'0801'), concat(year,'0901'), concat(year,'1001'), concat(year,'1101'), concat(year,'1201')])::date as month, unnest(array[sum(jan), sum(feb), sum(mar), sum(apr), sum(may), sum(jun), sum(jul), sum(aug), sum(sep), sum(oct), sum(nov), sum(dec)]) AS balance from fundbalancereport as fbr
left join account as a on a.accountnumber = split_part(text,' ',1) and a.organization = fbr.organization
where year > 2018
and left(text,1) = '1'
and accounttype = 'Bank'
group by fbr.organization, fbr.year, a.accounttype

这会产生我需要的确切格式，例如：

"organization","year","accounttype","month","balance"
"org1",2020,"Bank","2020-01-01","500.00000"
"org1",2020,"Bank","2020-02-01","550.00000"
"org1",2020,"Bank","2020-03-01","650.00000"
"org1",2020,"Bank","2020-04-01","450.00000"
"org1",2020,"Bank","2020-05-01","450.00000"
"org1",2020,"Bank","2020-06-01","450.00000"
"org1",2020,"Bank","2020-07-01","450.00000"
"org1",2020,"Bank","2020-08-01","450.00000"
"org1",2020,"Bank","2020-09-01","450.00000"
"org1",2020,"Bank","2020-10-01","450.00000"
"org1",2020,"Bank","2020-11-01","450.00000"
"org1",2020,"Bank","2020-12-01","450.00000"

理想情况下，我想添加一个 where 子句 of month is less than or equal to today 这应该产生：

"organization","year","accounttype","month","balance"
"org1",2020,"Bank","2020-01-01","500.00000"
"org1",2020,"Bank","2020-02-01","550.00000"
"org1",2020,"Bank","2020-03-01","650.00000"
"org1",2020,"Bank","2020-04-01","450.00000"

我看完了这个问题：

但是当我尝试添加额外的 SELECT 时，我收到一个错误，提示返回了多行，所以我有点不知所措，因为我已经有一个我需要的现有连接。

我试过了：

select fbr.organization, fbr.year, a.accounttype, 
(select unnest(array[concat(year,'0101'), concat(year,'0201'), concat(year,'0301'), concat(year,'0401'), concat(year,'0501'), concat(year,'0601'), concat(year,'0701'), concat(year,'0801'), concat(year,'0901'), concat(year,'1001'), concat(year,'1101'), concat(year,'1201')])::date as month), (select unnest(array[sum(jan), sum(feb), sum(mar), sum(apr), sum(may), sum(jun), sum(jul), sum(aug), sum(sep), sum(oct), sum(nov), sum(dec)]) AS balance) from fundbalancereport as fbr
left join account as a on a.accountnumber = split_part(text,' ',1) and a.organization = fbr.organization
where year > 2018
and left(text,1) = '1'
and accounttype = 'Bank'
group by fbr.organization, fbr.year, a.accounttype

得到了

ERROR:  more than one row returned by a subquery used as an expression

任何指导都会很有帮助，谢谢！

Answer 1

从 select 部分移动 unnest

select unnest(array['20190101'::date, '20200101'::date, '20210101'::date]) d

到查询的from部分并添加where部分

select d
from unnest(array['20190101'::date, '20200101'::date, '20210101'::date]) d
where d<=now()

编辑：

我相信你想要逆透视数据，在这种情况下你的查询可能是：

select fbr.organization, fbr.year, a.accounttype, fbr.month, sum(fbr.val) as balance
from
  (
    select organization, year, text, concat(year,'0101')::date as month, jan as val from fundbalancereport union
    select organization, year, text, concat(year,'0201')::date as month, feb as val from fundbalancereport union
    select organization, year, text, concat(year,'0301')::date as month, mar as val from fundbalancereport union
    select organization, year, text, concat(year,'0401')::date as month, apr as val from fundbalancereport union
    select organization, year, text, concat(year,'0501')::date as month, may as val from fundbalancereport union
    select organization, year, text, concat(year,'0601')::date as month, jun as val from fundbalancereport union
    select organization, year, text, concat(year,'0701')::date as month, jul as val from fundbalancereport union
    select organization, year, text, concat(year,'0801')::date as month, aug as val from fundbalancereport union
    select organization, year, text, concat(year,'0901')::date as month, sep as val from fundbalancereport union
    select organization, year, text, concat(year,'1001')::date as month, oct as val from fundbalancereport union
    select organization, year, text, concat(year,'1101')::date as month, nov as val from fundbalancereport union
    select organization, year, text, concat(year,'1201')::date as month, dec as val from fundbalancereport
  ) fbr
left join account as a on a.accountnumber = split_part(text,' ',1) and a.organization = fbr.organization
where year > 2018
and left(text,1) = '1'
and accounttype = 'Bank'
group by fbr.organization, fbr.year, a.accounttype, fbr.month

PostgresQL：在具有现有连接的 Unnest 上使用 Where 子句

PostgresSQL: Using Where Clause on Unnest with Existing Join

postgresql

join

unnest