我想使用union对查询中的值进行分组和删除

Question

Select
  pnf_notas_processadas.pnf_notas_dest_cnpj_cpf,
  Count(Distinct pnf_notas_processadas.pnf_notas_nnf) As destinadas,
  Sum(0) As emitidas
From
  pnf_notas_processadas
Group By
  pnf_notas_processadas.pnf_notas_dest_cnpj_cpf
Having
  pnf_notas_processadas.pnf_notas_dest_cnpj_cpf In ('03846642000102',
  '03846642000285', '03846642000447', '03846642000528')
Union
Select
  pnf_notas_processadas.pnf_notas_emit_cnpj_cpf,
  Sum(0) As destinadas,
  Count(Distinct pnf_notas_processadas.pnf_notas_nnf)
From
  pnf_notas_processadas
Group By
  pnf_notas_processadas.pnf_notas_emit_cnpj_cpf
Having
  pnf_notas_processadas.pnf_notas_emit_cnpj_cpf In ('03846642000102',
  '03846642000285', '03846642000447', '03846642000528')

结果：

我对查询的期望

Answer 1

一种方法是一步完成聚合。但是，您需要以某种方式聚合两个字段。

这是一种方法。基本上，行是＆＃34;加倍＆＃34;然后聚合拉出正确的值：

select (case when which = 1 then np.pnf_notas_dest_cnpj_cpf else pnf_notas_emit_cnpj_cpf end) as cnpj_cpf,
       count(distinct case when which = 1 then np.pnf_notas_nnf end) as destinadas,
       count(distinct case when which = 2 then np.pnf_notas_nnf end) as emitidas
from pnf_notas_processadas np cross join
     (values (1), (2)) as v(which)
where np.pnf_notas_dest_cnpj_cpf In ('03846642000102', '03846642000285', '03846642000447', '03846642000528')
Group By cnpj_cpf;

Answer 2

我调整了查询​​

select (case when which = 1 then  pnf_notas_processadas.pnf_notas_dest_cnpj_cpf else  pnf_notas_processadas.pnf_notas_emit_cnpj_cpf end) as cnpj_cpf,
       count(distinct case when which = 1 then  pnf_notas_processadas.pnf_notas_nnf end) as destinadas,
       count(distinct case when which = 2 then  pnf_notas_processadas.pnf_notas_nnf end) as emitidas
from  pnf_notas_processadas cross join
(values (1),  (2)) as v(wich)
where  pnf_notas_processadas.pnf_notas_dest_cnpj_cpf In ('03846642000102', '03846642000285', '03846642000447', '03846642000528')
Group By cnpj_cpf;

但是返回了此错误

选择句子使第5行，第16行的意外元素“2”无效