This may have been answered here before, but I have no clue how to search for it, so my apologies if it's a duplicate. I have a table with integer column, which holds number in sequence.
Some of numbers is missing due to removing of rows:
+-------+
| _auto |
+-------+
| 0 |
| 1 |
| 2 |
| 5 |
| 6 |
| 7 |
| 9 |
| 11 |
| 12 |
| 13 |
| 14 |
| 16 |
| 17 |
| 19 |
| 20 |
+-------+
What I want, is to write SQL query which gave me a table consists of ranges on this column.
+----+----+
| F | T |
+----+----+
| 0 | 2 |
| 5 | 7 |
| 9 | 9 |
| 11 | 14 |
| 16 | 17 |
| 19 | 20 |
+----+----+
I'll be using it with QT4::QSqlDatabase driver, so it would be nice if it's work on any compatible database (ANSI SQL), not only PostgreSQL.
I believe in this algorithm: (1) select rows, such that there is no rows with values _auto+1; (2) the same, but _auto-1; (3) union this two selects.
But I have no understanding how to write first two selects.
Sequences of numbers have a property: if you subtract an sequence from them, then the result is a constant. The following query uses this observation:
select min(_auto) as f, max(_auto) as t
from (select n._auto, row_number() over (order by n._auto) as seqnum
from numbers n
) n
group by (_auto - seqnum);