sql different columns compare - sql

Question

I have a database table with six columns(column1,column2...column6). The data are ordered and none repeat. These are the data in the database table
col1 col2 col3 col4 col5 col6
--------------------------------------------
1 3 4 6 7 8
2 5 7 9 10 14
I want to write sql to compare the data, if I have the data which one different number in varied/shift position. These are the parameters in the sql select statement.
col1 col2 col3 col4 col5 col6
--------------------------------------------
2 3 4 6 7 8
3 4 6 7 8 9
1 2 4 6 7 8
I want to query out the row 1,3,4,6,7,8
More complex...two different numbers and three different numbers
Merry Christmas!!
The following sql is from my friend.
data in table: Col1=10, Col2=11, Col3=12, Col4=13, Col5=26, Col6=28
parameters: 10,11,12,18,26,28
select * from
( select id,Col1,Col2, Col3,Col4, Col5,Col6,
( (case when Col1=10 then 1 else 0 end)
+(case when Col2=10 then 1 else 0 end)
+(case when Col3=10 then 1 else 0 end)
+(case when Col4=10 then 1 else 0 end)
+(case when Col5=10 then 1 else 0 end)
+(case when Col6=10 then 1 else 0 end)
)
+( (case when Col1=11 then 1 else 0 end)
+(case when Col2=11 then 1 else 0 end)
+(case when Col3=11 then 1 else 0 end)
+(case when Col4=11 then 1 else 0 end)
+(case when Col5=11 then 1 else 0 end)
+(case when Col6=11 then 1 else 0 end)
)
+( (case when Col1=12 then 1 else 0 end)
+(case when Col2=12 then 1 else 0 end)
+(case when Col3=12 then 1 else 0 end)
+(case when Col4=12 then 1 else 0 end)
+(case when Col5=12 then 1 else 0 end)
+(case when Col6=12 then 1 else 0 end)
)
+( (case when Col1=18 then 1 else 0 end)
+(case when Col2=18 then 1 else 0 end)
+(case when Col3=18 then 1 else 0 end)
+(case when Col4=18 then 1 else 0 end)
+(case when Col5=18 then 1 else 0 end)
+(case when Col6=18 then 1 else 0 end)
)
+( (case when Col1=26 then 1 else 0 end)
+(case when Col2=26 then 1 else 0 end)
+(case when Col3=26 then 1 else 0 end)
+(case when Col4=26 then 1 else 0 end)
+(case when Col5=26 then 1 else 0 end)
+(case when Col6=26 then 1 else 0 end)
)
+( (case when Col1=28 then 1 else 0 end)
+(case when Col2=28 then 1 else 0 end)
+(case when Col3=28 then 1 else 0 end)
+(case when Col4=28 then 1 else 0 end)
+(case when Col5=28 then 1 else 0 end)
+(case when Col6=28 then 1 else 0 end)
) as sub
from [DBName1].[dbo].[nList] ) aa
wheCole aa.sub>=5
Please give more comments and new answers!
Thanks for the reply !

Answer 1

My version of the solution (for SQL Server 2005 and above):
-- PREPARATIONS
create table #tbl1 (col1 int, col2 int, col3 int, col4 int, col5 int, col6 int);
create table #tbl2 (col1 int, col2 int, col3 int, col4 int, col5 int, col6 int);
insert into #tbl1 (col1, col2, col3, col4, col5, col6) values
(1, 3, 4, 6, 7, 8),
(2, 5, 7, 9, 10, 14);
insert into #tbl2 (col1, col2, col3, col4, col5, col6) values
(2, 3, 4, 6, 7, 8),
(3, 4, 6, 7, 8, 9),
(1, 2, 4, 6, 7, 8);
go
create function [dbo].[CompareDelimitedStrings] (#value1 varchar(max), #value2 varchar(max), #separator char(1))
returns int
as
begin
declare #result int = 0;
with r1 as
(
select value, cast(null as varchar(max)) [x], 0 [no] from (select rtrim(cast(#value1 as varchar(max))) [value]) as j
union all
select right(value, len(value)-case charindex(#separator, value) when 0 then len(value) else charindex(#separator, value) end) [value]
, left(r.[value], case charindex(#separator, r.value) when 0 then len(r.value) else abs(charindex(#separator, r.[value])-1) end ) [x]
, [no] + 1 [no]
from r1 r where value > ''
)
, r2 as
(
select value, cast(null as varchar(max)) [x], 0 [no] from (select rtrim(cast(#value2 as varchar(max))) [value]) as j
union all
select right(value, len(value)-case charindex(#separator, value) when 0 then len(value) else charindex(#separator, value) end) [value]
, left(r.[value], case charindex(#separator, r.value) when 0 then len(r.value) else abs(charindex(#separator, r.[value])-1) end ) [x]
, [no] + 1 [no]
from r2 r where value > ''
)
select #result = count(*)
from (
select x, [no] from r1 where x is not null
intersect
select x, [no] from r2 where x is not null
) as t
return #result;
end
go
-- SOLUTION
with [t1] as
(
select *
, replace(str(col1) + ',' + str(col2) + ',' + str(col3) + ',' + str(col4) + ',' + str(col5) + ',' + str(col6), ' ', '') [str]
from #tbl1
)
, [t2] as
(
select *
, replace(str(col1) + ',' + str(col2) + ',' + str(col3) + ',' + str(col4) + ',' + str(col5) + ',' + str(col6), ' ', '') [str]
from #tbl2
)
select distinct t1.col1, t1.col2, t1.col3, t1.col4, t1.col5, t1.col6
from t1
-- number 5 in this case means 5 intersections (or 1 difference).
-- you can change this number to 4 or 3 to find rows with 2 or 3 differences
join t2 on [dbo].[CompareDelimitedStrings] (t1.[str], t2.[str], ',') >= 5;
-- CLEANUP
drop table #tbl1;
drop table #tbl2;
drop function [dbo].[CompareDelimitedStrings];
Result:
col1 col2 col3 col4 col5 col6
--------------------------------------------
1 3 4 6 7 8
Algorithm in a few words: joining these two tables on the next condition: intersection of the numbers in the rows must be greater or equal to 5 (for the case with 1 difference)
This is the one-table-solution (shows the matched pairs and intersections count)
-- PREPARATIONS
create table #tbl (col1 int, col2 int, col3 int, col4 int, col5 int, col6 int);
insert into #tbl (col1, col2, col3, col4, col5, col6) values
(1, 3, 4, 6, 7, 8),
(2, 5, 7, 9, 10, 14),
(2, 3, 4, 6, 7, 8),
(3, 4, 6, 7, 8, 9),
(1, 2, 4, 6, 7, 8);
go
create function [dbo].[CompareDelimitedStrings] (#value1 varchar(max), #value2 varchar(max), #separator char(1))
returns int
as
begin
declare #result int = 0;
with r1 as
(
select value, cast(null as varchar(max)) [x], 0 [no] from (select rtrim(cast(#value1 as varchar(max))) [value]) as j
union all
select right(value, len(value)-case charindex(#separator, value) when 0 then len(value) else charindex(#separator, value) end) [value]
, left(r.[value], case charindex(#separator, r.value) when 0 then len(r.value) else abs(charindex(#separator, r.[value])-1) end ) [x]
, [no] + 1 [no]
from r1 r where value > ''
)
, r2 as
(
select value, cast(null as varchar(max)) [x], 0 [no] from (select rtrim(cast(#value2 as varchar(max))) [value]) as j
union all
select right(value, len(value)-case charindex(#separator, value) when 0 then len(value) else charindex(#separator, value) end) [value]
, left(r.[value], case charindex(#separator, r.value) when 0 then len(r.value) else abs(charindex(#separator, r.[value])-1) end ) [x]
, [no] + 1 [no]
from r2 r where value > ''
)
select #result = count(*)
from (
select x, [no] from r1 where x is not null
intersect
select x, [no] from r2 where x is not null
) as t
return #result;
end
go
-- SOLUTION
with [t] as
(
select *
, replace(str(col1) + ',' + str(col2) + ',' + str(col3) + ',' + str(col4) + ',' + str(col5) + ',' + str(col6), ' ', '') [str]
, row_number() over(order by col1, col2, col3, col4, col5, col6) [id]
from #tbl
)
, cross_dedup as
(
select t1.col1, t1.col2, t1.col3, t1.col4, t1.col5, t1.col6
, [dbo].[CompareDelimitedStrings] (t1.[str], t2.[str], ',') [intersections_count]
-- ranking the cross-doubled pairs
, row_number() over (partition by case when t1.id > t2.id then t1.id else t2.id end order by t1.id, t2.id) [rank]
from t t1
-- number 5 in this case means 5 intersections (or 1 difference).
-- you can change this number to 4 or 3 to find rows with 2 or 3 differences
join t t2 on [dbo].[CompareDelimitedStrings] (t1.[str], t2.[str], ',') >= 5
and t1.id != t2.id
)
select distinct col1, col2, col3, col4, col5, col6, [intersections_count] from cross_dedup where [rank] = 1;
-- CLEANUP
drop table #tbl;
drop function [dbo].[CompareDelimitedStrings];
Result:
col1 col2 col3 col4 col5 col6 intersections_count
-------------------------------------------------------------------
1 2 4 6 7 8 5
1 3 4 6 7 8 5