This is mainly a database connection issue, but this requires the use of the column-to-column mapping component.
I have several overview tables; for simplicity, let's say I have two, one since 2010 (called "d2010"), and the other since 2011 (called "d2011"). The rows correspond to the polls, and the columns (which we can consider named "c1", "c2", "c3", ... "cN") correspond to the answers to the questions. So, 'd2010' looks like
id c1 c2 c3 .. cX
survey 1
survey 2
survey 3
...
survey N
while 'd2011' looks like
id c1 c2 c3 .. cY
survey 1
survey 2
survey 3
...
survey N
- , 2010 2011 . , - SQL, ( ), , , :
id 2010-response 2011-response
survey 1
survey 2
survey 3
...
survey N
, , , ,
select d2011.id, d2010.c1 as 'last year', d2011.c1 'this year' from d2010 join d2011 on d2010.id=d2011.id;
, , , ; ,
c1 of 2010 corresponds to c2 of 2011
c2 of 2010 corresponds to c1 of 2011
c3 of 2010 corresponds to c3 of 2011
c4 of 2010 corresponds to c7 of 2011
...
, , -
select d2010.id, d2010.c1 as 'last year', d2011.f(c1) 'this year' from d2010 join d2011 on d2010.id=d2011.id;
"d2011.f(c1)" , 2010 2011 . , , -
2010 2011
c1 c2
c2 c1
c3 c3
c4 c7
SQL , ? , , SQL ( ) SQLite ( )?
select d2011.id, f(d2010.c1) as 'last year', d2011.c1 'this year' from d2010 join d2011 on d2010.id=d2011.id;.id;
!