SQL Double Dynamic Pivot - sql

Question

I am working on a double dynamic pivot based on 2 columns (HardwarePhase & HardwarePhase_Result).
Using the first result set in the image below, is the raw data that I have. Each set of 5 items (highlighted in images) are grouped based on HardwareTestCaseID.
The second result set in the image, is the current results that I'm getting from how I've constructed this query. Ideally, the result of the second column would be the same results, but instead it would be grouping the results.
The grouping is supposed to be based on the HardwareTestCaseID, but however, this is not happening.
The results I actually want are shown here. (There should be multiple rows, but this is just how it should be grouped per 5 entries).
This is the query I am currently using:
NOTE: The #col variables are built up based on the list of HardwarePhases (P0, M1, M2, M3).
select #query = 'SELECT ' + #colsNames + ',' + #colsResultNames + ', HardwareTestCaseID FROM
(
SELECT HardwarePhase_Result, HardwarePhase, ResultValue, HardwareTestCaseID, HardwareStatus
FROM #temp4
) as x
pivot
(
MAX(ResultValue)
FOR HardwarePhase_Result IN (' + #colsResult + ')
) as p
pivot
(
MAX(HardwareStatus)
FOR HardwarePhase IN (' + #cols + ')
) as p2 ';
using this table:
create table #temp4
(
HardwarePhase nvarchar(max),
HardwarePhase_Result nvarchar(max),
ResultValue bigint,
HardwareTestCaseID bigint,
HardwareStatus nvarchar(max),
Block nvarchar(max)
);

Answer 1

I personally would do it slightly different since you want to PIVOT on two columns. I would look at unpivoting the data in the multiple columns first, then apply the PIVOT function. I also would suggest that you start with writing a hard-coded version of the query first then convert it to dynamic SQL - this allows you to get the correct logic.
To unpivot the data, I would use CROSS APPLY so you can convert the pairs of columns into rows at the same time, the syntax would be similar to the following:
select col, value, HardwareTestCaseID
from temp4
cross apply
(
select HardwarePhase, HardwareStatus union all
select HardwarePhase_Result, cast(ResultValue as varchar(10))
) c (col, value)
See SQL Fiddle with Demo. Your data is then in the format:
| COL | VALUE | HARDWARETESTCASEID |
|-----------|-------------|--------------------|
| P0 | Not Started | 365 |
| P0_Result | 1 | 365 |
| M1 | Pass | 365 |
| M1_Result | 1 | 365 |
| M4 | Pass | 365 |
| M4_Result | 1 | 365 |
| M2 | Blocked | 365 |
| M2_Result | 1 | 365 |
Then you just apply the pivot function to the data:
select M1, M2, M3, M4, P0,
M1_Result, M2_Result, M3_Result,
M4_Result, P0_Result,
HardwareTestCaseID
from
(
select col, value, HardwareTestCaseID
from temp4
cross apply
(
select HardwarePhase, HardwareStatus union all
select HardwarePhase_Result, cast(ResultValue as varchar(10))
) c (col, value)
) d
pivot
(
max(value)
for col IN (M1, M2, M3, M4, P0,
M1_Result, M2_Result, M3_Result,
M4_Result, P0_Result)
) piv;
See SQL Fiddle with Demo.
Once you have the logic down, then convert it to dynamic SQL:
DECLARE #cols AS NVARCHAR(MAX),
#query AS NVARCHAR(MAX)
select #cols = STUFF((SELECT ',' + QUOTENAME(col)
from temp4
cross apply
(
select HardwarePhase, 1 union all
select HardwarePhase_Result, 2
) c (col, so)
group by col, so
order by so, col
FOR XML PATH(''), TYPE
).value('.', 'NVARCHAR(MAX)')
,1,1,'')
set #query = 'SELECT ' + #cols + ' , HardwareTestCaseID
from
(
select col, value, HardwareTestCaseID
from temp4
cross apply
(
select HardwarePhase, HardwareStatus union all
select HardwarePhase_Result, cast(ResultValue as varchar(10))
) c (col, value)
) x
pivot
(
max(value)
for col in (' + #cols + ')
) p '
execute sp_executesql #query;
See SQL Fiddle with Demo. This process gets a result:
| M1 | M2 | M3 | M4 | P0 | M1_RESULT | M2_RESULT | M3_RESULT | M4_RESULT | P0_RESULT | HARDWARETESTCASEID |
|---------|---------|---------|---------|-------------|-----------|-----------|-----------|-----------|-----------|--------------------|
| Pass | Blocked | Pass | Pass | Not Started | 1 | 1 | 1 | 1 | 1 | 365 |
| Blocked | Blocked | Blocked | Blocked | Pass | 1 | (null) | 1 | 1 | 1 | 366 |

Answer 2

--This is Just AWESOME. Simplified it for just one table as it's a much more common case (and could not find anything even close to this elegant after trying for hours)
DECLARE #cols AS NVARCHAR(MAX),
#query AS NVARCHAR(MAX)
select #cols = STUFF((SELECT ',' + QUOTENAME(your_key_column)
from YOUR_ORIGINAL_KEY_AND_VALUE_TABLE
group by your_key_column
FOR XML PATH(''), TYPE
).value('.', 'NVARCHAR(MAX)')
,1,1,'')
set #query = 'SELECT your_row_heading_columns,' + #cols + '
INTO YOUR_NEW_PIVOTED_TABLE
from
(
select your_row_heading_columns,your_key_column,your_value_column
from YOUR_ORIGINAL_KEY_AND_VALUE_TABLE
) x
pivot
(
max(your_value_column)
for your_key_column in (' + #cols + ')
) p '
execute sp_executesql #query;