Question

I have a postgres table that contains a very large number of numerals of various lengths. For example, the first column of the table will look like this:

1. 1
2. 12
3. 13
4. 134
5. 135
6. 136
7. 1362
8. 1363
9. …

I need postgres to return the row that matches as much of the beginning of a given numeral as possible. For example, the numeral “1358302” needs to match row 5 (“135”) and “1362304” needs to match row 7 (“1362”).

What SQL could I use to achieve this?

(the numerals aren’t sorted there permanently yet, so if you need to switch the data type between integer or string for simplicity/performance sake, I can still do that!)

Answer 1

SELECT n
FROM your_table
WHERE n IN (
    select $1/(10^i)::int8 from generate_series(0,floor(log($1))::int) i
)
ORDER BY n DESC
LIMIT 1