迁移到pgsql，select in tos in of select的函数中的错误

Question

我正在从sql-server迁移到pgsql。我使用Java中创建的脚本进行此操作，但是当我尝试将此过程（SQL-Server）转换为BD（PGSQL）中的功能和测试时，零件＃PASO1中的控制台显示错误

。

这是SQL-Server中的代码：

/****** Object:  StoredProcedure [dbo].[paBalanceClasificado]    Script Date: 30/11/2017 16:38:42 ******/
SET ANSI_NULLS OFF
GO
SET QUOTED_IDENTIFIER OFF
GO
CREATE PROCEDURE [dbo].[paBalanceClasificado]
@empresa int,
@fecha1 smalldatetime,
@fecha2 smalldatetime
AS
SELECT     SUBSTRING(b.codigoCuenta, 1, 1) + '000000' AS Codigo, SUM(a.montoDebe) AS Debe, SUM(a.montoHaber) AS Haber
INTO #PASO1
FROM         movimientosContables a INNER JOIN
                      CuentasContables b ON a.codigoCorto = b.codigoCorto AND b.codigoEmpresa = @empresa
WHERE     (a.codigoEmpresa = @empresa and fechaProceso between @fecha1 and @fecha2)
GROUP BY SUBSTRING(b.codigoCuenta, 1, 1) + '000000'
UNION ALL
SELECT     SUBSTRING(b.codigoCuenta, 1, 2) + '00000' AS Codigo, SUM(a.montoDebe) AS Debe, SUM(a.montoHaber) AS Haber
FROM         movimientosContables a INNER JOIN
                      CuentasContables b ON a.codigoCorto = b.codigoCorto AND b.codigoEmpresa = @empresa
WHERE     (a.codigoEmpresa = @empresa and fechaProceso between @fecha1 and @fecha2)
GROUP BY SUBSTRING(b.codigoCuenta, 1, 2) + '00000'
UNION ALL
SELECT     SUBSTRING(b.codigoCuenta, 1, 4) + '000' AS Codigo, SUM(a.montoDebe) AS Debe, SUM(a.montoHaber) AS Haber
FROM         movimientosContables a INNER JOIN
                      CuentasContables b ON a.codigoCorto = b.codigoCorto AND b.codigoEmpresa = @empresa
WHERE     (a.codigoEmpresa = @empresa and fechaProceso between @fecha1 and @fecha2)
GROUP BY SUBSTRING(b.codigoCuenta, 1, 4) + '000'
UNION ALL
SELECT     SUBSTRING(b.codigoCuenta, 1, 7) AS Codigo, SUM(a.montoDebe) AS Debe, SUM(a.montoHaber) AS Haber
FROM         movimientosContables a INNER JOIN
                      CuentasContables b ON a.codigoCorto = b.codigoCorto AND b.codigoEmpresa = @empresa
WHERE     (a.codigoEmpresa = @empresa and fechaProceso between @fecha1 and @fecha2)
GROUP BY SUBSTRING(b.codigoCuenta, 1,7)
ORDER BY codigo
SELECT     codigo, nombreCuenta, debe, haber,codigocorto
FROM         #Paso1, cuentascontables
WHERE     codigocuenta = codigo AND codigoempresa = @empresa --and 
SUBSTRING(codigoCuenta, 5,3)='000'
order by codigo
GO

这是PGSQL生成的代码（错误在变量＃paso1中，但我不知道该行是否存在更多错误。）：

CREATE  OR REPLACE FUNCTION pabalanceclasificado(
p_empresa int , 
p_fecha1 timestamp(0) , 
p_fecha2 timestamp(0) 
)
     RETURNS VOID
 AS $$
    DECLARE
    BEGIN
            select     substring(b.codigocuenta, 1, 1) + '000000' as codigo, sum(a.montodebe) as debe, sum(a.montohaber) as haber
        into #paso1
        from         movimientoscontables a inner join
                              cuentascontables b on a.codigocorto = b.codigocorto and b.codigoempresa = p_empresa
        where     (a.codigoempresa = p_empresa and fechaproceso between p_fecha1 and p_fecha2)
        group by substring(b.codigocuenta, 1, 1) + '000000'
        union all
        select     substring(b.codigocuenta, 1, 2) + '00000' as codigo, sum(a.montodebe) as debe, sum(a.montohaber) as haber
        from         movimientoscontables a inner join
                              cuentascontables b on a.codigocorto = b.codigocorto and b.codigoempresa = p_empresa
        where     (a.codigoempresa = p_empresa and fechaproceso between p_fecha1 and p_fecha2)
        group by substring(b.codigocuenta, 1, 2) + '00000'
        union all
        select     substring(b.codigocuenta, 1, 4) + '000' as codigo, sum(a.montodebe) as debe, sum(a.montohaber) as haber
        from         movimientoscontables a inner join
                              cuentascontables b on a.codigocorto = b.codigocorto and b.codigoempresa = p_empresa
        where     (a.codigoempresa = p_empresa and fechaproceso between p_fecha1 and p_fecha2)
        group by substring(b.codigocuenta, 1, 4) + '000'
        union all
        select     substring(b.codigocuenta, 1, 7) as codigo, sum(a.montodebe) as debe, sum(a.montohaber) as haber
        from         movimientoscontables a inner join
                              cuentascontables b on a.codigocorto = b.codigocorto and b.codigoempresa = p_empresa
        where     (a.codigoempresa = p_empresa and fechaproceso between p_fecha1 and p_fecha2)
        group by substring(b.codigocuenta, 1,7)
        order by codigo
        select     codigo, nombrecuenta, debe, haber,codigocorto
        from         #paso1, cuentascontables
        where     codigocuenta = codigo and codigoempresa = p_empresa --and substring(codigocuenta, 5,3)='000'
        order by codigo
END;
$$ LANGUAGE plpgsql;

控制台显示：

ERROR:  error de sintaxis en o cerca de «#»
LÍNEA 11:   into #paso1
> Terminated with exit code 0.

，但我认为这并不重要，因为我知道这不是做到这一点的方法...而且我不知道如何corectly

Answer 1

该功能中有多个错误：

1. 字符串串联使用SQL中的||进行。因此substring(b.codigocuenta, 1, 1) + '000000'应该是substring(b.codigocuenta, 1, 1) || '000000'
2. 标识符不允许从#开始。因此，#paso1是一个非法变量名称
3. 您没有声明变量#paso1。
4. 您不能将多行存储在一个变量中，因此看起来您想返回过程中查询的结果。
5. 要返回结果，您需要将功能声明为returns table而不是returns void
6. 要从PL/PGSQL函数返回结果集，您需要使用RETURN QUERY。您不能只写选择语句，也不对结果做任何事情。

但是您不需要PL/PGSQL来返回查询结果。SQL功能就足够了。

CREATE  OR REPLACE FUNCTION pabalanceclasificado(p_empresa int, p_fecha1 timestamp(0), p_fecha2 timestamp(0))
     RETURNS table (codigo varchar, debe bigint, bigint haber)
 AS $$
  select codigo, nombrecuenta, debe, haber,codigocorto
  from (
      select     substring(b.codigocuenta, 1, 1) || '000000' as codigo, sum(a.montodebe) as debe, sum(a.montohaber) as haber
      from  movimientoscontables a 
         inner join cuentascontables b on a.codigocorto = b.codigocorto and b.codigoempresa = p_empresa
      where     (a.codigoempresa = p_empresa and fechaproceso between p_fecha1 and p_fecha2)
      group by substring(b.codigocuenta, 1, 1) || '000000'
      union all
      select     substring(b.codigocuenta, 1, 2) || '00000' as codigo, sum(a.montodebe) as debe, sum(a.montohaber) as haber
      from         movimientoscontables a inner join
                            cuentascontables b on a.codigocorto = b.codigocorto and b.codigoempresa = p_empresa
      where     (a.codigoempresa = p_empresa and fechaproceso between p_fecha1 and p_fecha2)
      group by substring(b.codigocuenta, 1, 2) || '00000'
      union all
      select     substring(b.codigocuenta, 1, 4) || '000' as codigo, sum(a.montodebe) as debe, sum(a.montohaber) as haber
      from  movimientoscontables a 
         inner join cuentascontables b on a.codigocorto = b.codigocorto and b.codigoempresa = p_empresa
      where     (a.codigoempresa = p_empresa and fechaproceso between p_fecha1 and p_fecha2)
      group by substring(b.codigocuenta, 1, 4) || '000'
      union all
      select     substring(b.codigocuenta, 1, 7) as codigo, sum(a.montodebe) as debe, sum(a.montohaber) as haber
      from         movimientoscontables a 
         inner join cuentascontables b on a.codigocorto = b.codigocorto and b.codigoempresa = p_empresa
      where     (a.codigoempresa = p_empresa and fechaproceso between p_fecha1 and p_fecha2)
      group by substring(b.codigocuenta, 1,7)
   ) t
     join cuentascontables 
       on codigocuenta = codigo 
      and codigoempresa = p_empresa --and substring(codigocuenta, 5,3)='000'
   order by codigo;
$$ 
LANGUAGE sql;

要使用该功能：

select *
from pabalanceclasificado(42, current_timestamp, current_timestamp);