Reputation: 455
I have a column as below and the output i want is as given below this example. Col1 is a numeric datatype.
Col1 OutputCol1
1234 round to 2000
2300000 round to 3000000
456789.23 round to 500000
Always first digit+1. I could use the round function with negative value but it rounds to lower if the second number is less than 5.
Upvotes: 0
Views: 273
Reputation: 1270993
Well . . . you can add 1 to the first digit and then pad out with zeroes:
select rpad(cast(substr(col1, 1, 1) as int) + 1,
log(10, col1),
'0'
)
Upvotes: 1
Reputation: 95090
Use string maipulation. Take the first digit, add one, add trailing zeros. In case a digit is followed by zeros only (1000 or 10.00) don't add 1.
select col1,
case when nvl(to_number(substr(to_char(col1), 2)),0) = 0 then
to_number(rpad(substr(to_char(trunc(col1)), 1, 1), length(to_char(trunc(col1))), '0'))
else
to_number(rpad(to_char(to_number(substr(to_char(trunc(col1)), 1, 1)) + 1), length(to_char(trunc(col1))), '0'))
end as x,
to_number(substr(to_char(trunc(col1)), 2))
from
Upvotes: 1
Reputation: 23588
Here's a couple of ways - one mathematical, the other using manipulation of the string output of col1 in scientific notation:
WITH sample_data AS (SELECT 1234 col1 FROM dual UNION ALL
SELECT 2300000 col1 FROM dual UNION ALL
SELECT 456789.23 col1 FROM dual UNION ALL
SELECT -183 col1 FROM dual UNION ALL
SELECT 1000 col1 FROM dual UNION ALL
SELECT 0.392 col1 FROM dual)
SELECT col1,
ceil(col1 / order_of_magnitude_of_col1) * order_of_magnitude_of_col1 output1,
CEIL(to_number(SUBSTR(sn, 1, 4))) * POWER(10, to_number(SUBSTR(sn, 6))) output2
FROM (SELECT col1,
power(10, floor(LOG(10, abs(col1)))) order_of_magnitude_of_col1,
to_char(col1, 'fms0.0EEEE') sn
FROM sample_data);
COL1 OUTPUT1 OUTPUT2
---------- ---------- ----------
1234 2000 2000
2300000 3000000 3000000
456789.23 500000 500000
-183 -100 -100
1000 1000 1000
0.392 0.4 0.4
Upvotes: 0