Concatenate columns in Pivot View - sql

Question

DECLARE #DynamicPivotQuery AS NVARCHAR(MAX)
DECLARE #ColumnName AS NVARCHAR(MAX)
--Get distinct values of the PIVOT Column
SELECT #ColumnName= ISNULL(#ColumnName + ',','')
+ QUOTENAME(ACT_DESC)
FROM (SELECT DISTINCT ACT_DESC FROM tpc) AS desc
--Prepare the PIVOT query using the dynamic
SET #DynamicPivotQuery =
N'SELECT MDCODE, ' + #ColumnName + '
FROM tpc
PIVOT(MAX(ACTUAL_DATE)
FOR tpcIN (' + #ColumnName + ')) AS PVTTable'
--Execute the Dynamic Pivot Query
EXEC sp_executesql #DynamicPivotQuery
This query displays:
MDCODE | sample1 | sample2
--------------------------
123 | 1/2014 |
123 | | 2/2014
123 | | 3/2014
What I want is this:
MDCODE | sample1 | sample2
--------------------------
123 | 1/2014 | 2/2014,3/2014
Does anyone have an idea how to concat the column data? Any tips?
This is the table where i get the data:
mdcode | act_desc | actual_date
--------------------------
1234 | sample1 | 1/2014
1234 | sample2 | 2/2014
1234 | sample2 | 3/2014
the actual_date is datetime

Answer 1

In SQL Server, the type of comma separated output is achieved with "for xml path", but is also usually associated with stuff()
Optionally via "cross apply" although it doesn't have to be done that way.
I would approach it like this. Don't use pivot now, just reduce MDCODE & ACT_DESC by distinct or group by, then using cross apply concatenate the string of "dates". If they really are dates then you will need to convert them to varchar e.g. convert(varchar,actual_date,112)
| MDCODE | ACT_DESC | ACTUAL_DATES |
|--------|----------|---------------|
| 1234 | sample1 | 1/2014 |
| 1234 | sample2 | 2/2014,3/2014 |
produced by:
SELECT
d.mdcode
, d.act_desc
, ca1.actual_dates
FROM (
SELECT DISTINCT
[mdcode]
, [act_desc]
FROM Table1
) d
CROSS APPLY (
SELECT
STUFF((
SELECT
',' + a.actual_date
FROM table1 a
WHERE a.mdcode = d.mdcode
AND a.act_desc = d.act_desc --<< change here
ORDER BY a.actual_date
FOR xml PATH ('')
)
, 1, 1, '')
) AS ca1 (actual_dates)
;
By the way STUFF() is used to remove the first comma from the concatenation, that is its only role. CROSS APPLY is arguably better than using a correlated subquery in the select clause, because it is performed before that clause as part of the from clause. As I said earlier its optional but I prefer it.
see: http://sqlfiddle.com/#!3/a24ba/4
& also see these samples

Answer 2

I would skip the PIVOT altogether and just use FOR XML_PATH, like this:
DECLARE #DynamicPivotQuery AS NVARCHAR(MAX)
--Get distinct values of the PIVOT Column
SELECT #DynamicPivotQuery= ISNULL(#DynamicPivotQuery + ',','select mdcode,')
+ 'stuff((select '',''+actual_date from tpc where mdcode=t.mdcode and act_desc = ''' + ACT_DESC + ''' for xml path(''''),type).value(''.'',''varchar(max)''),1,1,'''') '
+ QUOTENAME(ACT_DESC)
FROM (SELECT DISTINCT ACT_DESC FROM tpc) AS des
select #DynamicPivotQuery = #DynamicPivotQuery + 'from tpc t group by mdcode'
EXEC sp_executesql #DynamicPivotQuery
The dynamic query generates a query like this:
select mdcode,
stuff(
(
select ','+actual_date
from tpc where mdcode=t.mdcode and act_desc = 'sample1'
for xml path(''),type
).value('.','varchar(max)')
,1,1,'') sample1,
stuff(
(
select ','+actual_date
from tpc where mdcode=t.mdcode and act_desc = 'sample2'
for xml path(''),type
).value('.','varchar(max)')
,1,1,'') sample2
from tpc t
group by mdcode;
The SQL Fiddle demonstrates the static query and the dynamic one