Question

I’m looking to ‘flatten’ my dataset in order to facilitate data mining.
Each categorical column should be changed to multiple Boolean columns.
I have a column with categorical values, e.g.:

 ID    col1
  1     A
  2     B
  3     A

I’m looking for a way to pivot this table, and have an aggregated function telling me whether this ID has value A or B:

Result:

 ID    col1A    col1B
  1     1        0
  2     0        1
  3     1        0

I tried using PIVOT but have no idea which aggregated function to use inside it.

Also looked for answers in SF but couldn’t find any…

I’m using MS-SQL 2012.

Any help would be appreciated!
Omri

EDIT:

The number of categories in col1 is unknown, therefore the solution must be dynamic.
Thanks 🙂

Answer 1

try this:

select ID,
         col1A=(case when col1='A' then 1 else 0 end),
         col1B=(case when col1='B' then 1 else 0 end)
  from <table>    

IF you have one ID with both A and B and you want to have distinct ID in the output you could do

 select ID,
         col1A=max(case when col1='A' then 1 else 0 end),
         col1B=max(case when col1='B' then 1 else 0 end)
  from <table> 
  group by id

EDIT

As per your comment, If you do not know the number of options for col1, then you can go for dynamic PIVOT

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

select @cols = STUFF((SELECT distinct ',' + QUOTENAME(col1) 
                    from <table> 
            FOR XML PATH(''), TYPE
            ).value('.', 'NVARCHAR(MAX)') 
        ,1,1,'')

set @query = 'SELECT id, ' + @cols + ' from <table> 

            pivot 
            (
                count([col1])
                for col1 in (' + @cols + ')
            ) p '
print(@query)
execute(@query)

SQL Fiddle Demo