Trouble with joining based on various combinations of columns

Question

Hi
I have a load of transaction data which I need to join to a rate table to find the correct rate to price on. The join should be one-to-one and the selection of the correct rate is based on various columns in both tables which should either match or, where
the rate table contains 'NA', allow any value from the transaction table. Where there is a specific match, this should be selected rather than a join where 'NA' is used to allow any other value in the transaction.
Here is the CREATE script for the transaction table:
CREATE TABLE [dbo].[job_variables](
[job_no] [int] NOT NULL,
[product_code1] [nvarchar](2) NULL,
[product_code2] [nvarchar](2) NULL,
[customer_code] [nvarchar](8) NULL,
[area1] [nvarchar](8) NULL,
[area2] [nvarchar](8) NULL,
[group_code] [nvarchar](8) NULL)
And for the rate table:
CREATE TABLE [dbo].[ratecriteria](
[product_code1] [nvarchar](255) NULL,
[product_code2] [nvarchar](255) NULL,
[customer_code] [nvarchar](255) NULL,
[area1] [nvarchar](255) NULL,
[area2] [nvarchar](255) NULL,
  [group_code] [nvarchar](255) NULL,)
The combinations used to join the tables should be attempted in this order, where X = a match and NA = ratecriteria value is 'NA' so allow all from job_variables. As the data tables need to join one-to-one, once a transaction meets the join criteria, it
cannot be joined to another ratecriteria record.
Firstly, where ratecriteria.group_code = 'NA' and customer_key matches:
product_code1
product_code2
area1
area2
customer_code
group_code
X
X
X
X
X
NA
X
X
X <------>
X
X
NA
(area1 and area2 reciprocated)
X
X
X
NA
X
NA
X
X
NA
X
X
NA
X
X
NA
NA
X
NA
NA
X
X
X
X
NA
NA
X
X
NA
X
NA
NA
X
NA
X
X
NA
NA
X
NA
NA
X
NA
Then, where ratecriteria.customer_code = 'NA' and group_code matches:
product_code1
area1
area2
customer_code
group_code
X
X
X
NA
X
X
X <------>
X
NA
X
(area1 and area2 reciprocated)
X
X
NA
NA
X
X
NA
X
NA
X
X
NA
NA
NA
X
NA
X
X
NA
X
NA
X
NA
NA
X
NA
NA
X
NA
X
NA
NA
NA
NA
X
And finally, where both ratecriteria.customer_code and ratecriteria.group_code = 'NA':
product_code1
area1
area2
customer_code
group_code
X
X
X
NA
NA
X
X <------>
X
NA
NA
(area1 and area2 reciprocated)
X
X
NA
NA
NA
X
NA
X
NA
NA
X
NA
NA
NA
NA
NA
X
X
NA
NA
NA
X
NA
NA
NA
NA
NA
X
NA
NA
NA
NA
NA
NA
NA
I am working with SQL Server 2012 in SSMS. Any help greatly appreciated, let me know if I need to provide more info.
Thanks

Answer 1

I think that you have got that correct, although I am not entirely sure what "REFERENCES" does, I will research that one. I do know that the database design is poor but, for context, I have imported all the data I am going to use, it
is all in excellent shape and I will not do any further imports, creates, deletes, updates etc. so a poor schema should (and this may be me showing my lack of knowledge again) not matter at this stage. I am really looking for a solution to the above problem.
In terms of a query which outputs a correctly selected rate code (see above post 5th March 15:01), my solution which is perhaps a sledgehammer to crack a nut is as such.
I have created a table I have called ratecrit_junction which I use to generate all the different combinations of actual values from job_variables and 'NA' entries.
CREATE TABLE [dbo].[ratecrit_junction](
[selection_rank] INT IDENTITY(1,1),
[product_code1] [char](2),
[product_code2] [char](2),
[area1] [varchar](6),
[area2] [varchar](6),
[customer_code] [varchar](8),
[group_code] [varchar](10)
INSERT INTO ratecrit_junction
VALUES ('X', 'X', 'X', 'X', 'X', 'NA'),
('X', 'X', 'X', 'NA', 'X', 'NA'),
('X', 'X', 'NA', 'X', 'X', 'NA'),
('X', 'X', 'NA', 'NA', 'X', 'NA'),
('NA', 'X', 'X', 'X', 'X', 'NA'),
('NA', 'X', 'X', 'NA', 'X', 'NA'),
('NA', 'X', 'NA', 'X', 'X', 'NA'),
('NA', 'X', 'NA', 'NA', 'X', 'NA'),
('X', 'X', 'X', 'X', 'NA', 'X'),('X', 'X', 'X', 'NA', 'NA', 'X'),
('X', 'X', 'NA', 'X', 'NA', 'X'),
('X', 'X', 'NA', 'NA', 'NA', 'X'),
('NA', 'X', 'X', 'X', 'NA', 'X'),
('NA', 'X', 'X', 'NA', 'NA', 'X'),
('NA', 'X', 'NA', 'X', 'NA', 'X'),
('NA', 'X', 'NA', 'NA', 'NA', 'X'),
('X', 'NA', 'X', 'X', 'X', 'NA'),
('X', 'NA', 'X', 'NA', 'X', 'NA'),
('X', 'NA', 'NA', 'X', 'X', 'NA'),
('X', 'NA', 'NA', 'NA', 'X', 'NA'),
('NA', 'NA', 'X', 'X', 'X', 'NA'),
('NA', 'NA', 'X', 'NA', 'X', 'NA'),
('NA', 'NA', 'NA', 'X', 'X', 'NA'),
('NA', 'NA', 'NA', 'NA', 'X', 'NA'),
('X', 'NA', 'X', 'X', 'NA', 'X'),
('X', 'NA', 'X', 'NA', 'NA', 'X'),
('X', 'NA', 'NA', 'X', 'NA', 'X'),
('X', 'NA', 'NA', 'NA', 'NA', 'X'),
('NA', 'NA', 'X', 'X', 'NA', 'X'),
('NA', 'NA', 'X', 'NA', 'NA', 'X'),
('NA', 'NA', 'NA', 'X', 'NA', 'X'),
('NA', 'NA', 'NA', 'NA', 'NA', 'X')
I can then join data from job_variables to ratecrit_junction using a COALESCE(NULLIF(...), ...) to get the different combinations and join to ratecriteria with a ranking to indicate which one of them should override the others and identify the correct rate
code to use.
I forgot to include rate code in my original post, correction below.
CREATE TABLE [dbo].[ratecriteria](
ID INT IDENTITY(1,1) PRIMARY KEY,
[product_code1] [nvarchar](2) REFERENCES Product_Code,
[product_code2] [nvarchar](2) REFERENCES Product_Code,
[customer_code] [nvarchar](8) REFERENCES Customer_Code,
[area1] [nvarchar](8) REFERENCES Area,
[area2] [nvarchar](8) REFERENCES Area,
[group_code] [nvarchar](8) REFERENCES Group_Code,
[rate_code] [nvarchar](5),
UNIQUE (product_code1, product_code2, customer_code, group_code)
And here is the query that I am using
SELECT v.job_no,
j.selection_rank,
rc.rate_code
FROM job_variables v
CROSS JOIN ratecrit_junction j
INNER JOIN ratecriteria rc
ON COALESCE(NULLIF(j.product_code1, 'X'), v.product_code1) = rc.product_code1
AND COALESCE(NULLIF(j.product_code2, 'X'), v.product_code2) = rc.product_code1
AND COALESCE(NULLIF(j.area1, 'X'), v.area1) = rc.area1
AND COALESCE(NULLIF(j.area2, 'X'), v.area2) = rc.area2
AND COALESCE(NULLIF(j.customer_code, 'X'), v.customer_code) = rc.customer_code
AND COALESCE(NULLIF(j.group_code, 'X'), c.group_code) = rc.group_code
And here is some sample data for the job_variables and ratecriteria tables, based on my earlier post with the coloured examples.
INSERT INTO ratecriteria
VALUES ('P1', 'PP1', 'C1', 'AA', 'BB', 'NA', 'R1'),
('P1', 'PP1', 'C1', 'AA', 'NA', 'NA', 'R2'),
('P1', 'PP1', 'NA', 'AA', 'BB', 'G1', 'R3'),
('NA', 'NA', 'NA', 'NA', 'NA', 'G1', 'R4')
INSERT INTO job_variables
VALUES (1, 'P1', 'PP1', 'C1', 'AA', 'BB', 'G3'),
(2, 'P1', 'PP1', 'C1', 'BB', 'AA', 'G4'),
(3, 'P1', 'PP1', 'C1', 'AA', 'FF', 'G1'),
(4, 'P1', 'PP1', 'C2', 'AA', 'BB', 'G1'),
(5, 'P1', 'PP2', 'C1', 'AA', 'BB', 'G1'),
(6, 'P2', 'PP2', 'C4', 'AB', 'AA', 'G1')