Question

我有一张包含以下记录的表格：

Id  Partitioned SDDate
1   1   2014-01-01
2   2   2014-01-02
3   3   2014-01-03
4   4   2014-01-04
5   5   2014-01-05
6   6   2014-01-06
7   7   2014-01-07
8   1   2014-01-08
9   2   2014-01-09
10  3   2014-01-10
11  4   2014-01-11
12  5   2014-01-12
13  6   2014-01-13
14  7   2014-01-14
15  1   2014-01-15
16  2   2014-01-16

正如您所看到的，记录是从1-7分区的，我想要实现的是将该组的记录插入临时表。达到最大分区ID数时，再次启动插入。感谢。

我正在使用SQL SERVER 2008 RC2

这是代码：

Set NoCount On;

--
Declare 
    @InputTable As DT_HorasExtras

--
Insert Into @InputTable
( HorasExtra, Fecha )
Values
(3, '2014-01-01'), 
(3, '2014-01-02'), 
(1, '2014-01-03'), 
(2, '2014-01-03'), 
(1, '2014-01-05'), 
(3, '2014-01-06'), 
(3, '2014-01-07'), 
(3, '2014-01-08'), 
(3, '2014-01-09'), 
(3, '2014-01-10'), 
(3, '2014-01-11'), 
(3, '2014-01-12'),
(3, '2014-01-13'),
(3, '2014-01-14'),
(3, '2014-01-15'),
(3, '2014-01-16')


Declare @Idx Int = 1, @InnerIncrementor Int = 1;

--
Declare 
    --
    @PartitionedId Int,
    @Fecha Date,
    @HorasExtra Int,
    @Id Int;

    --
    DECLARE IDs CURSOR LOCAL FOR
        Select 
            t.Id
            From utilsFilterAndPartitionHorasExtras(@InputTable) t

    OPEN IDs
    FETCH NEXT FROM IDs into @Id
        WHILE @@FETCH_STATUS = 0
        BEGIN

            --
            Select 
                @HorasExtra = t.HorasExtra,
                @Fecha = t.Fecha,
                @PartitionedId = t.Partitioned
                From utilsFilterAndPartitionHorasExtras(@InputTable) t
                    Where t.Id = @Id

            --
            print 'Particion Dia: ' + Convert(varchar(max), @PartitionedId) + ', Fecha: ' + Convert(varchar(max), @Fecha) + ', Horas extras: ' + Convert(varchar(max), @HorasExtra);

        --*** HERE I WANT TO INSERT RECORDS ONLY FROM PARTITION 1 TO 7
        --*** WHEN 7 IS REACHED START INSERT AGAIN
         insert into #tmpTable values(@Field1,.,,


        -- inc
        Set @Idx+=1;
    FETCH NEXT FROM IDs into @Id
    END
    CLOSE IDs
    DEALLOCATE IDs

这是输出：

Particion Dia: 1, Fecha: 2014-01-01, Horas extras: 3
Particion Dia: 2, Fecha: 2014-01-02, Horas extras: 3
Particion Dia: 3, Fecha: 2014-01-03, Horas extras: 3
Particion Dia: 4, Fecha: 2014-01-05, Horas extras: 1
Particion Dia: 5, Fecha: 2014-01-06, Horas extras: 3
Particion Dia: 6, Fecha: 2014-01-07, Horas extras: 3
Particion Dia: 7, Fecha: 2014-01-08, Horas extras: 3
Particion Dia: 1, Fecha: 2014-01-09, Horas extras: 3
Particion Dia: 2, Fecha: 2014-01-10, Horas extras: 3
Particion Dia: 3, Fecha: 2014-01-11, Horas extras: 3
Particion Dia: 4, Fecha: 2014-01-12, Horas extras: 3
Particion Dia: 5, Fecha: 2014-01-13, Horas extras: 3
Particion Dia: 6, Fecha: 2014-01-14, Horas extras: 3
Particion Dia: 7, Fecha: 2014-01-15, Horas extras: 3
Particion Dia: 1, Fecha: 2014-01-16, Horas extras: 3

Answer 1

如果要为这些行分配组ID，则可以从id中减去分区：

select (partition - id) as grp, t.*
from table t

如果id值确实不一定是连续的，您可以将row_number()用于此目的：

select (row_number() over () order by (id, partition) - id) as grp, t.*
from table t

这将识别每个组。我不确定你插入它们并重新开始是什么意思。

迭代表并在分区行上执行操作？

1 个答案: