Pivoting tables in MS SQL SERVER - sql-server

Question

I have some data that I want to pivot.
---------------------------------------------
| date | price | col_1 | col_2 |
---------------------------------------------
| 2017-12-10 | 26 | fruit | apple |
| 2017-12-10 | 346 | Vege | carrot |
| 2017-12-11 | 644 | Sweet | cake |
| 2017-12-11 | 35435 | fruit | banana |
| 2017-12-12 | 453455 | veggie| beans |
---------------------------------------------
col_1 and col_2 are combinations of categories
So basically the output should be 1 row per unique date, and the rest of the columns should be relevant combinations of col_1 and col2 separated by a hyphen. The table will be populated by the total price of the combination of col_1 col_2 item per date. the diagram below might be useful:
-------------------------------------------------------------------
| date | fruit - apple| fruit - banana| val n-SubVal n |
-------------------------------------------------------------------
| 2017-12-10 | NULL | 56 | and so on
| 2017-12-11 | 100 | NULL |
| 2017-12-12 | NULL | 900 |
| 2017-12-13 | 45 | NULL |
| 2017-12-14 | NULL | NULL |
--------------------------------------------------------------------
after about 4 hours talking to a friend and research we have come up with this script below. It executes but nothing gets returned except a very odd error.
Incorrect syntax near '*some value in either col_1 or col_2*'.
the error is supposedly happening either at the first or second declare statement (whichever line MSSQL randomly chooses as its never the same line #).
Any help in getting the above output will be appreciated.
Thank you very much.
declare #dynamicpivotquery as NVARCHAR(MAX)
declare #columnname as NVARCHAR(MAX)
select #columnname = COALESCE(#columnname + ', ','') + quotename(col_1) + quotename(col_1)
from (select distinct col_1, col_2 from *tablename*) as d
set #dynamicpivotquery =
N'select distinct date, ' + #columnname + ' from *tablename*
PIVOT (sum(price) for d in (' + #columnname + ')) as pivot12'
EXEC sp_executesql #dynamicpivotquery

Answer 1

Some remarks on your query:
The sample data has fieldnames date, price, col_1 and col_2 while your query uses fieldnames of sold_date, sold_price, cat_1 and cat_2.
You could have had a look at the result in the variable #dynamicpivotquery to see what's wrong: [fruit][apple] is not a valid fieldname. It should have been [fruit - apple] instead.
The PIVOT query did not know the field d with the concatenated category names.
The distinct in the dynamic query isn't needed. You have to make sure that only the required fields (sold_date, sold_price and d) are contained in the source for the PICOT command, because otherwise query will also be grouped by all additional fields, which will result in unwanted rows.
Your code could work like this:
DECLARE #dynamicpivotquery nvarchar(MAX)
DECLARE #columnnames nvarchar(MAX)
SELECT #columnnames =
ISNULL(#columnnames + ', ', '') + QUOTENAME(ISNULL(cat_1, '') + ' - ' + ISNULL(cat_2, ''))
FROM (SELECT DISTINCT cat_1, cat_2 FROM tablename) AS d
SET #dynamicpivotquery =
N'SELECT sold_date, ' + #columnnames + N'
FROM
(SELECT sold_date, sold_price,
ISNULL(cat_1, '''') + '' - '' + ISNULL(cat_2, '''') AS d
FROM tablename
) AS src
PIVOT (SUM(sold_price) FOR d IN (' + #columnnames + ')) AS pivot12'
EXEC sp_executesql #dynamicpivotquery

Answer 2

You can try the following query :
declare #dynamicpivotquery as NVARCHAR(MAX)
declare #columnname as NVARCHAR(MAX)
select #columnname = COALESCE(#columnname + ', ','') + QuoteName(cat)
from (select distinct cat_1+'-'+cat_2 as cat from #YourTable) as d
SET #dynamicpivotquery =
N';WITH p AS (
SELECT sold_date, [cat_1] +''-'' +[cat_2] AS CATCOL, SUM(sold_price) AS sold_price
FROM #YourTable
GROUP BY sold_date, [cat_1] + ''-'' + [cat_2]
)
SELECT sold_date, ' + #columnname + '
FROM p
PIVOT (SUM([sold_price]) FOR CATCOL IN (' + #columnname + ')) AS pivotcat12'
EXEC sp_executesql #dynamicpivotquery
Sample data :
sold_date sold_price cat_1 cat_2 cat_12
----------------------------------------------------
2017-12-10 26,00 fruit apple fruit-apple
2017-12-10 346,00 vege carrot vege-carrot
2017-12-11 644,00 sweet cake sweet-cake
2017-12-11 35435,00 fruit banana fruit-banana
2017-12-12 453455,00 veggie beans veggie-beans
2017-12-12 100,00 other fruits other-fruits
2017-12-12 100,00 other fruits other-fruits
2017-12-12 100,00 other fruits other-fruits
Dynamic query string :
;WITH p AS (
SELECT sold_date, [cat_1] +'-' +[cat_2] AS CATCOL, SUM(sold_price) AS sold_price
FROM #YourTable
GROUP BY sold_date, [cat_1] + '-' + [cat_2]
)
SELECT sold_date, [fruit-apple], [fruit-banana], [other-fruits], [sweet-cake], [vege-carrot], [veggie-beans]
FROM p
PIVOT (
SUM([sold_price])
FOR CATCOL IN ([fruit-apple], [fruit-banana], [other-fruits], [sweet-cake], [vege-carrot], [veggie-beans])) AS pivotcat12
Result :
sold_date fruit-apple fruit-banana other-fruits sweet-cake vege-carrot veggie-beans
--------------------------------------------------------------------------------------------
2017-12-10 26,00 NULL NULL NULL 346,00 NULL
2017-12-11 NULL 35435,00 NULL 644,00 NULL NULL
2017-12-12 NULL NULL 300,00 NULL NULL 453455,00

Answer 3

Pivot The data Twice. Just apply the same logic dynamically so that it appears for all combinations of col_1 and col_2.
select * from
( select [DATE], [col_1] ,[col_2],[price]
from *tablename*
) TableName
pivot
(sum(price)
for col_1 in (fruit)
) as FirstPivot
pivot
(sum(fruit)
for col_2 in (apple,banana)
) as SecondPivot