SQL Pivot在这种情况下？

Question

I'm trying to make this

DATE             Insurance  Type    Billed  Rate
2/28/2011 0:00   BC/BS      INP     B       0.6383
2/28/2011 0:00   BC/BS      OUT     B       0.5216
2/28/2011 0:00   BC/BS      INP     U       0.1988
2/28/2011 0:00   BC/BS      OUT     U       0.3493
3/31/2011 0:00   BC/BS      INP     B       0.69
3/31/2011 0:00   BC/BS      OUT     B       0.6136
3/31/2011 0:00   BC/BS      INP     U       0.1877
3/31/2011 0:00   BC/BS      OUT     U       0.3567

Look like this

Insurance   Type    Billed  2/28/2011 0:00  3/31/2011 0:00
BC/BS       INP     B       0.6383           0.69
BC/BS       OUT     B       0.5216           0.6136
BC/BS       INP     U       0.1988           0.1877
BC/BS       OUT     U       0.3493           0.3567

So that the date field row data ends up becoming the column headers for each distinct value. I think I can utilize the PIVOT statement but all of the examples I look at seem to be to simple for this.

Thanks in advance.

Answer 1

Based on the sample data that you provided, I am guessing you will have a long list of dates that you will need to turn into columns. While you can use a static pivot like the other answer, you can use a dynamic pivot similar to below:

DECLARE @cols AS NVARCHAR(MAX),
    @query  AS NVARCHAR(MAX)

select @cols = STUFF((SELECT distinct ',' + QUOTENAME(convert(char(10), date, 101)) 
                    from test
            FOR XML PATH(''), TYPE
            ).value('.', 'NVARCHAR(MAX)') 
        ,1,1,'')

set @query 
      = 'SELECT Insurance, [Type], Billed,' + @cols + ' from 
         (
            SELECT [date], 
               Insurance, 
               [Type], 
               Billed, 
               Rate 
            FROM test
         ) x
         pivot 
         (
            sum(Rate)
            for [date] in (' + @cols + ')
         ) p 
         order by billed'

execute(@query)

See SQL Fiddle with Demo

Answer 2

If the pivot columns (your Dates) are known then you can use the PIVOT operator to get your results:

declare @t table (date datetime, Insurance varchar(10), Type char(3), Billed char(1), Rate decimal(10,4));

    insert into @t
        values  ('2/28/2011 0:00',   'BC/BS',      'INP',     'B',       '0.6383'),
                ('2/28/2011 0:00',   'BC/BS',      'OUT',     'B',       '0.5216'),
                ('2/28/2011 0:00',   'BC/BS',      'INP',     'U',       '0.1988'),
                ('2/28/2011 0:00',   'BC/BS',      'OUT',     'U',       '0.3493'),
                ('3/31/2011 0:00',   'BC/BS',      'INP',     'B',       '0.69'),
                ('3/31/2011 0:00',   'BC/BS',      'OUT',     'B',       '0.6136'),
                ('3/31/2011 0:00',   'BC/BS',      'INP',     'U',       '0.1877'),
                ('3/31/2011 0:00',   'BC/BS',      'OUT',     'U',       '0.3567')

    select  *
    from    (   select  [date] as pivot_col, 
                        Insurance, 
                        [Type], 
                        Billed, 
                        Rate 
                from    @t
            ) as d
    pivot   (   sum(Rate) for pivot_col in ([2/28/2011],[3/31/2011])
            ) as p;

Its likely you dont know the values and if that's the case you will need to look into using the same technique but in a dynamic way. A great example is here