> I think the Oracle guy's version could easily be adapted to PG 8.4 ---
> those little rownum subqueries seem to be just a substitute for not
> having generate_series(1,9), and everything else is just string-pushing.
indeed.
marcin=# with recursive x( s, ind ) as
( select sud, position( ' ' in sud )
from (select '53 7 6 195 98 6 8 6 34 8 3 17 2
6 6 28 419 5 8 79'::text as sud) xx
union all
select substr( s, 1, ind - 1 ) || z || substr( s, ind + 1 )
, position(' ' in repeat('x',ind) || substr( s, ind + 1 ) )
from x
, (select gs::text as z from generate_series(1,9) gs)z
where ind > 0
and not exists ( select null
from generate_series(1,9) lp
where z.z = substr( s, ( (ind - 1 ) / 9 ) * 9 + lp, 1 )
or z.z = substr( s, mod( ind - 1, 9 ) - 8 + lp * 9, 1 )
or z.z = substr( s, mod( ( ( ind - 1 ) / 3 ), 3 ) * 3
+ ( ( ind - 1 ) / 27 ) * 27 + lp
+ ( ( lp - 1 ) / 3 ) * 6
, 1 )
)
)
select s
from x
where ind = 0;
s
-----------------------------------------------------------------------------------
534678912672195348198342567859761423426853791713924856961537284287419635345286179
(1 row)
Time: 472.027 ms
btw: it is pretty cool to replace some of the numbers in input with
spaces and see how the query finds multiple solutions
btw2: is SQL with 'with recursive' turing-complete ? Anyone care to
try a Brainf*ck interpreter ? :)
Greetings
marcin Mańk