3 table join

Question

Hi
Lets say we have following tables
TabA
colA1 (PK)
colA2
colA3
TabB
colB1 (PK)
colA1 (FK - Tab A)
colB2
TabC
colC1 (PK)
colB1 (FK - Tab B)
colA1 (FK - Tab A)
colC2when linking above three tables what is the difference between following two scripts
select * from TabA a, TabB b, TabC c
where a.colA1 = b.colA1 and b.colB1 = c.colB1
select * from TabA a, TabB b, TabC c
where a.colA1 = b.colA1 and b.colB1 = c.colB1 and c.colA1 = a.colA1What is the usage of both
Cheers
Shabar

Answer 1

Hi Shabar
Worked with the sample data.With tow joins data is correct.i.e. it gives us 5 rows
with ft_books as
(select 2800 BK_CODE,'PL/SQL' BK_NAME,'Wembly' BK_PUBLISHER ,'Computer' BK_CATEGORY, 250 AU_CODE from dual
union all
select 2600, 'Data Mining','Wilkinson','Computer',250 from dual
union all
select 1575, 'Deadly Creatures','Williams','Wild Life',150 from dual
union all
select 1500, 'Modern Robotics','Samson','Science',350 from dual
union all
select  2500, 'Database Design','Wilkinson','Computer',250 from dual),
ft_authors as
select 150 AU_CODE,'Wilson'  AU_NAME,'Peters Avenue, New York' AU_ADDS,'USA' AU_COUNTRY from dual
union all
select 350,'Martin Clark','Will street, Paris','France' from dual
union all
select 250,'James Blake','90, 3rd street, London','England' from dual),
ft_book_lons as
select 112 BL_CODE,2800 BL_BK_CODE,20 BL_MEM_ID,'12/08/2010' BL_BWR_DATE,'23/09/2010' BL_RTN_DATE,250 BL_AU_CODE from dual
union all
select  182,2600,33,'12/07/2010','20/09/2010',250 from dual
union all
select 189,1500,33,'12/07/2010','20/09/2010',350 from dual
union all
select 199,1500,19,'06/08/2010','11/10/2010',350 from dual
union all
select 506,1575,13,'06/02/2000','18/07/2000',null from dual
select * from ft_books a,ft_authors b,ft_book_lons c
where a.au_code=b.au_code
and a.bk_code=c.bl_bk_code
Output is
2800     PL/SQL     Wembly     Computer     250     250     James Blake     90, 3rd street, London     England     112     2800     20     12/08/2010     23/09/2010     250
2600     Data Mining     Wilkinson     Computer     250     250     James Blake     90, 3rd street, London     England     182     2600     33     12/07/2010     20/09/2010     250
1500     Modern Robotics     Samson     Science     350     350     Martin Clark     Will street, Paris     France     189     1500     33     12/07/2010     20/09/2010     350
1500     Modern Robotics     Samson     Science     350     350     Martin Clark     Will street, Paris     France     199     1500     19     06/08/2010     11/10/2010     350Now when you add third condition.In that case as the value in ft_book_lons.BL_CODE=506 has bl_au_code as null so when it is comparing with other table value its returning false & hence depreciates one row.
You can modify your query in case of third condiition as below
with ft_books as
(select 2800 BK_CODE,'PL/SQL' BK_NAME,'Wembly' BK_PUBLISHER ,'Computer' BK_CATEGORY, 250 AU_CODE from dual
union all
select 2600, 'Data Mining','Wilkinson','Computer',250 from dual
union all
select 1575, 'Deadly Creatures','Williams','Wild Life',150 from dual
union all
select 1500, 'Modern Robotics','Samson','Science',350 from dual
union all
select  2500, 'Database Design','Wilkinson','Computer',250 from dual),
ft_authors as
select 150 AU_CODE,'Wilson'  AU_NAME,'Peters Avenue, New York' AU_ADDS,'USA' AU_COUNTRY from dual
union all
select 350,'Martin Clark','Will street, Paris','France' from dual
union all
select 250,'James Blake','90, 3rd street, London','England' from dual),
ft_book_lons as
select 112 BL_CODE,2800 BL_BK_CODE,20 BL_MEM_ID,'12/08/2010' BL_BWR_DATE,'23/09/2010' BL_RTN_DATE,250 BL_AU_CODE from dual
union all
select  182,2600,33,'12/07/2010','20/09/2010',250 from dual
union all
select 189,1500,33,'12/07/2010','20/09/2010',350 from dual
union all
select 199,1500,19,'06/08/2010','11/10/2010',350 from dual
union all
select 506,1575,13,'06/02/2000','18/07/2000',null from dual
select * from ft_books a,ft_authors b,ft_book_lons c
where a.au_code=b.au_code
and a.bk_code=c.bl_bk_code
and nvl(b.au_code,0)=nvl(c.bl_au_code,0);
Output is
2800     PL/SQL     Wembly     Computer     250     250     James Blake     90, 3rd street, London     England     112     2800     20     12/08/2010     23/09/2010     250
2600     Data Mining     Wilkinson     Computer     250     250     James Blake     90, 3rd street, London     England     182     2600     33     12/07/2010     20/09/2010     250
1500     Modern Robotics     Samson     Science     350     350     Martin Clark     Will street, Paris     France     189     1500     33     12/07/2010     20/09/2010     350
1500     Modern Robotics     Samson     Science     350     350     Martin Clark     Will street, Paris     France     199     1500     19     06/08/2010     11/10/2010     350Hope it helps!!!
Thanks
AJ