如何在特定条件下合并行 postgres Sql

Question

我有一个问题。下面是我的 sql.

SELECT upniki.naziv,
        (
            SELECT count(caseid) 
            FROM terjatve, paket 
            WHERE terjatve.stevilkapaketa=paket.id_paket
            AND terjatve.idnarocnik=upniki.id_upnik
            AND paket.datum_predaje >='1900-03-01'
            AND paket.datum_predaje <='2020-03-31' 
            GROUP BY upniki.naziv) as st_vseh_primerov,
        (
            SELECT sum(racuni.startsum) 
            FROM racuni, paket, terjatve 
            WHERE terjatve.stevilkapaketa=paket.id_paket 
            AND terjatve.caseid=racuni.caseid
            AND terjatve.idnarocnik=upniki.id_upnik 
            AND paket.datum_predaje >='1900-03-01'
            AND paket.datum_predaje <='2020-03-31' 
            GROUP BY upniki.naziv) as glavnica_vseh_primerov,
        (
            SELECT count(caseid) 
            FROM terjatve, paket 
            WHERE terjatve.stevilkapaketa=paket.id_paket 
            AND terjatve.idnarocnik=upniki.id_upnik 
            AND paket.datum_predaje >='1900-03-01' 
            AND paket.datum_predaje <='2020-03-31'
            AND terjatve.statusnacinid='1' 
            GROUP BY upniki.naziv) as st_aktivnih
FROM upniki 
GROUP BY upniki.naziv, upniki.id_upnik

有了这个 SQL 我得到的结果如下。

naziv     st_vseh_primerov  glavnica_vseh_primerov  st_aktivnih
CLIENT 1       12                      7500                  0
CLIENT 2       10                     40000                  0
CLIENT 3       15                      5000                  0
CLIENT 4       16                     15000                  0
CLIENT 5        9                     12000                  0
CLIENT 6        8                     60000                  1
CLIENT 7        4                      3000                  0
CLIENT 8        3                     10000                  0

我需要的结果是将 st_vseh_primerov <=10 的行合并为一行。作为组合主题的行将不再显示。所以，结果应该是这样的：

naziv      st_vseh_primerov     glavnica_vseh_primerov      st_aktivnih
CLIENT 1          12                            7500              0
CLIENT 3          15                            5000              0
CLIENT 4          16                           15000              0
SME               34                          125000              1

如果有人能帮忙，请。

Br R.

Answer 1

您可以将现有查询转换为 cte 或子查询，并在外部查询中聚合：

with cte as (... your query here ...)
select
    case when c.st_vseh_primerov <= 10 
        then 'SME'
        else c.naviz
    end naviz,
    SUM(st_vseh_primerov) st_vseh_primerov,
    SUM(glavnica_vseh_primerov) glavnica_vseh_primerov,
    SUM(st_aktivnih) st_aktivnih
from cte c
group by
    case when c.st_vseh_primerov <= 10 
        then 'SME'
        else c.naviz
    end

请注意，您不应使用隐式连接（在 from 子句中使用逗号）；这是几十年前的旧语法，不应该出现在新代码中。相反，您应该使用标准的显式连接（使用 on 关键字）。所以基本上，像这样的东西（来自你的第一个子查询）：

FROM terjatve, paket 
WHERE 
    terjatve.stevilkapaketa=paket.id_paket
    AND terjatve.idnarocnik=upniki.id_upnik

应该写成：

FROM terjatve
INNER JOIN paket ON terjatve.stevilkapaketa=paket.id_paket
WHERE terjatve.idnarocnik=upniki.id_upnik

Answer 2

你可以用它的案例。

select case when ( st_vseh_primerov is null or st_vseh_primerov <= 10 ) then 'SME' else naziv end as naziv,
       sum(st_vseh_primerov) as st_vseh_primerov,
       sum(glavnica_vseh_primerov) as glavnica_vseh_primerov,
       max(case when ( st_vseh_primerov is null or st_vseh_primerov <= 10 ) then 1 else 0 end) as st_aktivnih
from ( < your select above ) tmp
group by case when ( st_vseh_primerov is null or st_vseh_primerov <= 10 ) then 'SME' else naziv end;

编辑：您也可以使用 coalesce()。即：

select case when coalesce( st_vseh_primerov,0) <= 10  then 'SME' else naziv end as naziv,
       sum(st_vseh_primerov) as st_vseh_primerov,
       sum(glavnica_vseh_primerov) as glavnica_vseh_primerov,
       max(case when coalesce( st_vseh_primerov,0) <= 10 then 1 else 0 end) as st_aktivnih
from ( < your select above ) tmp
group by case when coalesce( st_vseh_primerov,0) <= 10 then 'SME' else naziv end;

如何在特定条件下合并行 postgres Sql

How to combine rows under certain conditions postgres Sql

sql

postgresql

group-by

rows

count