Can someone break down a PIVOT query for me please? - sql

Question

I am looking to pivot this data so the output can count how many people are in each age band...
Below is what I have done, but I don't even think I'm close....
I am using SQL Server 2008
SELECT [OA_Code], Dominant_F_Age, Age_F_90plus, Age_F_85_89, Age_F_80_84, Age_F_75_79, Age_F_70_74, Age_F_65_69, Age_F_60_64,
Age_F_55_59, Age_F_50_54, Age_F_45_49, Age_F_40_44, Age_F_35_39, Age_F_30_34, Age_F_25_29, Age_F_20_24, Age_F_15_19, Age_F_10_14, Age_F_5_9,
Age_F_0_4, Age_2001_F
FROM cen.AgeByGenderOA
PIVOT (
f_Age_data
for [F_Age_Data] in (Age_F_90plus, Age_F_85_89, Age_F_80_84, Age_F_75_79, Age_F_70_74, Age_F_65_69, Age_F_60_64,
Age_F_55_59, Age_F_50_54, Age_F_45_49, Age_F_40_44, Age_F_35_39, Age_F_30_34, Age_F_25_29, Age_F_20_24, Age_F_15_19, Age_F_10_14, Age_F_5_9),
Age_F_0_4)
My table looks like so:
CREATE TABLE AgeByGenderOA(
[OA_Code] AS VARCHAR(50),
[Age_M_0_4] [varchar](50) NULL,
[Age_M_5_9] [varchar](50) NULL,
[Age_M_10_14] [varchar](50) NULL,
[Age_M_15_19] [varchar](50) NULL,
[Age_M_20_24] [varchar](50) NULL,
[Age_M_25_29] [varchar](50) NULL,
[Age_M_30_34] [varchar](50) NULL,
[Age_M_35_39] [varchar](50) NULL,
[Age_M_40_44] [varchar](50) NULL,
[Age_M_45_49] [varchar](50) NULL,
[Age_M_50_54] [varchar](50) NULL,
[Age_M_55_59] [varchar](50) NULL,
[Age_M_60_64] [varchar](50) NULL,
[Age_M_65_69] [varchar](50) NULL,
[Age_M_70_74] [varchar](50) NULL,
[Age_M_75_79] [varchar](50) NULL,
[Age_M_80_84] [varchar](50) NULL,
[Age_M_85_89] [varchar](50) NULL,
[Age_M_90plus] [varchar](50) NULL,
[Dominant_M_Age] [varchar](50) NULL,
[Age_2001_F] [varchar](50) NULL,
[Age_F_0_4] [varchar](50) NULL,
[Age_F_5_9] [varchar](50) NULL,
[Age_F_10_14] [varchar](50) NULL,
[Age_F_15_19] [varchar](50) NULL,
[Age_F_20_24] [varchar](50) NULL,
[Age_F_25_29] [varchar](50) NULL,
[Age_F_30_34] [varchar](50) NULL,
[Age_F_35_39] [varchar](50) NULL,
[Age_F_40_44] [varchar](50) NULL,
[Age_F_45_49] [varchar](50) NULL,
[Age_F_50_54] [varchar](50) NULL,
[Age_F_55_59] [varchar](50) NULL,
[Age_F_60_64] [varchar](50) NULL,
[Age_F_65_69] [varchar](50) NULL,
[Age_F_70_74] [varchar](50) NULL,
[Age_F_75_79] [varchar](50) NULL,
[Age_F_80_84] [varchar](50) NULL,
[Age_F_85_89] [varchar](50) NULL,
[Age_F_90plus] [varchar](50) NULL,
[Dominant_F_Age] [varchar](50) NULL,
[MAPINFO_ID] [varchar](50) NULL)
As you can see it's a huge table. The "Age_M_0_4" columns are populated with numbers.
I would like my output to read
OA_Code, Age, countOfAge
123456, Age_0_4, 26
123456, Age_5_9, 24
789456, Age_0_4, 10
789456, Age_5_9, 12
This would be inclusive of male and female ages.
Anyone fancy taking a crack at this one?
Finally: The data contained in the "Age_M_" Columns is all numbers counts. 50, 40, 2, 0 etc...

Answer 1

You do not want a PIVOT, you need to UNPIVOT the data. Since you are using SQL Server 2008+, you can use CROSS APPLY with a VALUES clause. The code will be similar to this:
select t.OA_Code,
c.age,
sum(cast(c.value as int)) countOfAge
from AgeByGenderOA t
cross apply
(
values
('Age_0_4', Age_M_0_4),
('Age_5_9', Age_M_5_9),
('Age_10_14', Age_M_10_14),
('Age_15_19', Age_M_15_19) ... add the other columns here
) c (age, value)
group by t.OA_Code, c.age
See SQL Fiddle with Demo
Edit, If you want both Male/Female columns you will include ALL of the columns in the values clause, but give them the same age range value:
select t.OA_Code,
c.age,
sum(cast(c.value as int)) countOfAge
from AgeByGenderOA t
cross apply
(
values
('Age_0_4', Age_M_0_4),
('Age_5_9', Age_M_5_9),
('Age_10_14', Age_M_10_14),
('Age_15_19', Age_M_15_19),
('Age_0_4', Age_F_0_4),
('Age_5_9', Age_F_5_9),
('Age_10_14', Age_F_10_14),
('Age_15_19', Age_F_15_19)
) c (age, value)
group by t.OA_Code, c.age;
See SQL Fiddle with Demo