SQL Server: What is the difference between CROSS JOIN and FULL OUTER JOIN?

A cross join produces a cartesian product between the two tables, returning all possible combinations of all rows. It has no on clause because you're just joining everything to everything.

A full outer join is a combination of a left outer and right outer join. It returns all rows in both tables that match the query's where clause, and in cases where the on condition can't be satisfied for those rows it puts null values in for the unpopulated fields.

This wikipedia article explains the various types of joins with examples of output given a sample set of tables.

One thing that might not always be obvious to some is that a cross join with an empty table (or result set) results in empty table (M x N; hence M x 0 = 0)

A full outer join will always have rows unless both M and N are 0.

I'd like to add one important aspect to other answers, which actually explained this topic to me in the best way:

If 2 joined tables contain M and N rows, then cross join will always produce (M x N) rows, but full outer join will produce from MAX(M,N) to (M + N) rows (depending on how many rows actually match "on" predicate).

EDIT:

From logical query processing perspective, CROSS JOIN does indeed always produce M x N rows. What happens with FULL OUTER JOIN is that both left and right tables are "preserved", as if both LEFT and RIGHT join happened. So rows, not satisfying ON predicate, from both left and right tables are added to the result set.

For SQL Server, CROSS JOIN and FULL OUTER JOIN are different. CROSS JOIN is simply Cartesian Product of two tables, irrespective of any filter criteria or any condition.

FULL OUTER JOIN gives unique result set of LEFT OUTER JOIN and RIGHT OUTER JOIN of two tables. It also needs ON clause to map two columns of tables.

Table 1 contains 10 rows and Table 2 contains 20 rows with 5 rows matching on specific columns. Then CROSS JOIN will return 10*20=200 rows in result set. FULL OUTER JOIN will return 25 rows in result set. INNER JOIN will return matching rows, hence, 5 rows in result set. FULL OUTER JOIN (or any other JOIN) always returns result set with less than or equal to Cartesian Product number. Number of rows returned by FULL OUTER JOIN equal to (No. of Rows by LEFT OUTER JOIN) + (No. of Rows by RIGHT OUTER JOIN) - (No. of Rows by INNER JOIN).

They are the same concepts, apart from the NULL value returned.

See below:

declare @table1 table( col1 int, col2 int );
insert into @table1 select 1, 11 union all select 2, 22;

declare @table2 table ( col1 int, col2 int );
insert into @table2 select 10, 101 union all select 2, 202;

select
    t1.*,
    t2.*
from @table1 t1
full outer join @table2 t2 on t1.col1 = t2.col1
order by t1.col1, t2.col1;

/* full outer join
col1        col2        col1        col2
----------- ----------- ----------- -----------
NULL        NULL        10          101
1           11          NULL        NULL
2           22          2           202
*/

select
    t1.*,
    t2.*
from @table1 t1
cross join @table2 t2
order by t1.col1, t2.col1;

/* cross join
col1        col2        col1        col2
----------- ----------- ----------- -----------
1           11          2           202
1           11          10          101
2           22          2           202
2           22          10          101
*/

Cross join :Cross Joins produce results that consist of every combination of rows from two or more tables. That means if table A has 3 rows and table B has 2 rows, a CROSS JOIN will result in 6 rows. There is no relationship established between the two tables – you literally just produce every possible combination.

Full outer Join : A FULL OUTER JOIN is neither "left" nor "right"— it's both! It includes all the rows from both of the tables or result sets participating in the JOIN. When no matching rows exist for rows on the "left" side of the JOIN, you see Null values from the result set on the "right." Conversely, when no matching rows exist for rows on the "right" side of the JOIN, you see Null values from the result set on the "left."

SQL FULL OUTER JOIN

The FULL OUTER JOIN returns all rows from the left table (table1) and from the right table (table2) irrespective of the match.

The FULL OUTER JOIN keyword combines the result of both LEFT OUTER JOIN and RIGHT OUTER JOIN

SQL full outer join is also known as FULL JOIN

Reference : http://datasciencemadesimple.com/sql-full-outer-join/

SQL CROSS JOIN

In SQL CROSS JOIN Each Row of first table is mapped with the each and every row of second table.

Number of rows produced by a result set of CROSS JOIN operation is equal to number of rows in the first table multiplied by the number of rows in the second table.

CROSS JOIN is also known as Cartesian product / Cartesian join

Number of rows in table A is m, Number of rows in table B is n and resultant table will have m*n rows

Reference:http://datasciencemadesimple.com/sql-cross-join/

Cross Join: http://www.dba-oracle.com/t_garmany_9_sql_cross_join.htm

TLDR; Generates a all possible combinations between 2 tables (Carthesian product)

(Full) Outer Join: http://www.w3schools.com/Sql/sql_join_full.asp

TLDR; Returns every row in both tables and also results that have the same values (matches in CONDITION)

A full outer join combines a left outer join and a right outer join. The result set returns rows from both tables where the conditions are met but returns null columns where there is no match.

A cross join is a Cartesian product that does not require any condition to join tables. The result set contains rows and columns that are a multiplication of both tables.

Want to add intuition on the common types of joins. Think of parametrized join pseudo-operator:

function generalJoin(
    forceIncludeLeft: boolean, forceIncludeRight: boolean, condition: Expression,
    tableA: Table, tableB: Table
) {
  for (leftRow in tableA) {
    for (rightRow in tableB) {
      // Output combined row if condition is true
      if (condition(leftRow, rightRow)) {
        markRowHasMatch(leftRow);
        markRowHasMatch(rightRow);
        output(leftRow.append(rightRow));
      }
    }
  }
  if (forceIncludeLeft) {
    for (leftRow in tableA) {
      if (!rowHasMatch(leftRow)) {
        output(leftRow.append(createArrayOfNulls(tableB.columnsCount)));
      }
    }
  }
  if (forceIncludeRight) {
    for (rightRow in tableB) {
      if (!rowHasMatch(rightRow)) {
        output(createArrayOfNulls(tableA.columnsCount).append(rightRow));
      }
    }
  }
}
// We may think of Expression as function taking rowA and rowB and
// returning boolean true or false
function alwaysTrue(rowA, rowB) {
  return true;
}

which is equivalent to

[SELECT ...] from tableA SOMEJOIN tableB ON condition

Then this is the equivalence table:

a [INNER] JOIN b ON condition       <=> generalJoin(false, false, condition, a, b)
a LEFT [OUTER] JOIN b ON condition  <=> generalJoin(true,  false, condition, a, b)
a RIGHT [OUTER] JOIN b ON condition <=> generalJoin(false, true,  condition, a, b)
a FULL [OUTER] JOIN b ON condition  <=> generalJoin(true,  true,  condition, a, b)
a CROSS JOIN b                      <=> generalJoin(false, false, alwaysTrue, a, b)

Effectively, CROSS JOIN is equivalent to INNER JOIN with no condition, i.e. every left row does match every right row.