December 22, 2016 at 10:19 am
I have a table with 2 columns, FormulaID, IngredientID
The table contains formulanames with the ingredients.
I want to find formula pairs within this table with the same ingredients.
Say the values are
FormulaID IngredientID
F1 I1
F1 I2
F2 I3
F2 I4
F2 I5
F3 I1
F3 I2
F4 I3
F4 I4
F4 I5
F5 I1
F5 I2
and so on
When I run the query, I want
F1 F3
F1 F5
F3 F5
F2 F4
and so on returned
I don't mind if it returns all combinations also, so result could be
F1 F3
F1 F5
F3 F1
F3 F5
F5 F1
F5 F3
F2 F4
I am stymied here, tried different things but not even close to the result
December 22, 2016 at 11:02 am
IF OBJECT_ID('tempdb..#tmp', 'U') IS NOT NULL
DROP TABLE #tmp;
CREATE TABLE #tmp
(
FormulaId CHAR(2)
, IngredientId CHAR(2)
);
INSERT #tmp
(FormulaId, IngredientId)
VALUES
('F1' ,'I1'),
('F1' ,'I2'),
('F2' ,'I3'),
('F2' ,'I4'),
('F2' ,'I5'),
('F3' ,'I1'),
('F3' ,'I2'),
('F4' ,'I3'),
('F4' ,'I4'),
('F4' ,'I5'),
('F5' ,'I1'),
('F5' ,'I2')
SELECT DISTINCT
t1.FormulaId
, t2.FormulaId
FROM
#tmp t1
JOIN #tmp t2 ON t1.IngredientId = t2.IngredientId
WHERE t2.FormulaId > t1.FormulaId;
The absence of evidence is not evidence of absence
- Martin Rees
The absence of consumable DDL, sample data and desired results is, however, evidence of the absence of my response
- Phil Parkin
December 22, 2016 at 11:20 am
Phil Parkin (12/22/2016)
IF OBJECT_ID('tempdb..#tmp', 'U') IS NOT NULLDROP TABLE #tmp;
CREATE TABLE #tmp
(
FormulaId CHAR(2)
, IngredientId CHAR(2)
);
INSERT #tmp
(FormulaId, IngredientId)
VALUES
('F1' ,'I1'),
('F1' ,'I2'),
('F2' ,'I3'),
('F2' ,'I4'),
('F2' ,'I5'),
('F3' ,'I1'),
('F3' ,'I2'),
('F4' ,'I3'),
('F4' ,'I4'),
('F4' ,'I5'),
('F5' ,'I1'),
('F5' ,'I2')
SELECT DISTINCT
t1.FormulaId
, t2.FormulaId
FROM
#tmp t1
JOIN #tmp t2 ON t1.IngredientId = t2.IngredientId
WHERE t2.FormulaId > t1.FormulaId;
This will give combos where they share ANY ingredient, but I think that he is looking for records where they have ALL of the same ingredients.
I've added another formula to the sample data that illustrates the problem. F6 should not match any of the other formulas, but it is matching each of the other formulas.
IF OBJECT_ID('tempdb..#tmp', 'U') IS NOT NULL
DROP TABLE #tmp;
CREATE TABLE #tmp
(
FormulaId CHAR(2)
, IngredientId CHAR(2)
);
INSERT #tmp
(FormulaId, IngredientId)
VALUES
('F1' ,'I1'),
('F1' ,'I2'),
('F2' ,'I3'),
('F2' ,'I4'),
('F2' ,'I5'),
('F3' ,'I1'),
('F3' ,'I2'),
('F4' ,'I3'),
('F4' ,'I4'),
('F4' ,'I5'),
('F5' ,'I1'),
('F5' ,'I2'),
('F6' ,'I1'),
('F6' ,'I5')
Drew
J. Drew Allen
Business Intelligence Analyst
Philadelphia, PA
December 22, 2016 at 11:32 am
This approach will work, but it may not be the most efficient. It uses XML concatenation to create a comma-separated list of the ingredients and then checks for two formulae that have the same comma-separated list.
;
WITH formulae AS (
SELECT DISTINCT t.FormulaId, STUFF((
SELECT ',', t2.IngredientId
FROM #tmp t2
WHERE t2.FormulaId = t.FormulaId
ORDER BY t2.IngredientId
FOR XML PATH(''), TYPE
).value('.', 'VARCHAR(255)'), 1, 1, '') AS IngredientList
FROM #tmp t
)
SELECT a.FormulaId, b.FormulaId
FROM formulae a
INNER JOIN formulae b
ON a.FormulaId < b.FormulaId
AND a.IngredientList = b.IngredientList
Drew
J. Drew Allen
Business Intelligence Analyst
Philadelphia, PA
December 22, 2016 at 12:22 pm
Drew,Thanks for your reply. I see that you have sorted the ingredients and this might work. let me try it out.
December 22, 2016 at 12:55 pm
hamzajosh (12/22/2016)
Drew,Thanks for your reply. I see that you have sorted the ingredients and this might work. let me try it out.
I'd also be interested to hear why my solution is not suitable, as it returns exactly the data you were looking for.
The absence of evidence is not evidence of absence
- Martin Rees
The absence of consumable DDL, sample data and desired results is, however, evidence of the absence of my response
- Phil Parkin
December 22, 2016 at 1:02 pm
Phil, Like Drew pointed out, your solution returns formulas that have even ONE ingredient in common, while I am looking for a exact match on ALL ingredients. I appreciate your effort and time. Drew, your solution worked. Awesome! thanks a ton.
December 22, 2016 at 1:23 pm
hamzajosh (12/22/2016)
Phil, Like Drew pointed out, your solution returns formulas that have even ONE ingredient in common, while I am looking for a exact match on ALL ingredients. I appreciate your effort and time. Drew, your solution worked. Awesome! thanks a ton.
OK, that was not clear from your original post (at least, not to me). Thanks for responding.
The absence of evidence is not evidence of absence
- Martin Rees
The absence of consumable DDL, sample data and desired results is, however, evidence of the absence of my response
- Phil Parkin
December 22, 2016 at 1:40 pm
Yes, My bad, sorry for not being clear.
Viewing 9 posts - 1 through 8 (of 8 total)
You must be logged in to reply to this topic. Login to reply