There are two tables with one-to-one relationship, and it’s a legal that the second table can have less rows than the first.
Here is a sql code with inserted data:
create table data( data_id int primary key, name text ); create table data_extra( extra_id int references data(data_id) primary key , extra text ); insert into data values (1, 'n1'); insert into data values (2, 'n2'); insert into data values (3, 'n3'); insert into data_extra values (1, 'e1'); insert into data_extra values (2, 'e2'); select * from data left join data_extra on data_id = extra_id; -- data_id | name | extra_id | extra -- ---------+------+----------+------- -- 1 | n1 | 1 | e1 -- 2 | n2 | 2 | e2 -- 3 | n3 | | -- (3 rows)
I need to make a query for all data
+ data_extra
tables where extra!=’e1′
Here is my try:
select * from data left join data_extra on data_id = extra_id where extra != 'e1'; -- data_id | name | extra_id | extra -- ---------+------+----------+------- -- 2 | n2 | 2 | e2 -- (1 row)
The problem with this query, that it doesn’t return data.name=’n3′ row, which doesn’t have a related row in data_extra table.
How can I fix it, to get as result data.name=n2 and data.name=n3.
Advertisement
Answer
You could do:
select * from data d left join data_extra de on d.data_id = de.extra_id where de.extra_id is null or de.extra <> 'e1'
This removes rows that match and whose extra
is equal to 'e1'
, while retaining non-matching rows.