combination of unique values a column in a table - sql

Question

I need to find the all unique possible combination of values in a column in a table. For example, for column values 1,2,3,4,5. i want the result to be [1,2],[1,3],[1,4],[1,5],[2,1],[2,3] etc.
Will appreciate any pointers to construct the query to find the combination of the values.
thanks

Answer 1

You can do a cross join in BigQuery by using a subselect that adds a constant key value, then joining on that constant value.
For example, here is a query that will compute the cross join of {1, 2, 3} and {2, 4, 6}:
SELECT t1.num as first, t2.num as second
FROM (
SELECT num, 1 as key
FROM (
SELECT 1 as num), (
SELECT 2 as num), (
SELECT 3 as num)) as t1
JOIN (
SELECT num, 1 as key
FROM (
SELECT 2 as num), (
SELECT 4 as num), (
SELECT 6 as num)) as t2
ON t1.key = t2.key
WHERE t1.num <> t2.num
Note this uses a BigQuery "trick" to create the two input tables. If you were just doing this with an existing table, it would look like:
SELECT t1.num as first, t2.num as second
FROM (
SELECT foo as num, 1 as key
FROM [my_dataset.my_table]) as t1
JOIN (
SELECT foo as num, 1 as key
FROM [my_dataset.my_table]) as t2
ON t1.key = t2.key
WHERE t1.num <> t2.num

Answer 2

A cross join might be usefull.
See this demo: http://www.sqlfiddle.com/#!12/59af5/1
The ANSI SQL syntax uses a CROSS JOIN operator:
create table val( x int );
insert into val values(1),(2),(3),(4),(5);
SELECT a.x a, b.x b
FROM val a
CROSS JOIN val b
WHERE a.x <> b.x
ORDER BY a,b;
Another form of this query without CROSS JOIN should work on most DBMS system, but ANSI form is recommended for clearness:
SELECT a.x a, b.x b
FROM val a, val b
WHERE a.x <> b.x
ORDER BY a,b;
Beware that the cross join for large datasets can kill your database performance, for 100 values it generates 100x100 = 10.000 rows, for 1000 --> 1.000.000 rows.