Reputation: 9
Im trying to come up with a test that validates decimals in a particular column (with 220000 records). For example for column A there shouldn't be any values with more decimals than 2, 1 is also ok.
for example :
Column A (datatype varchar)
48528.64
135082.54
5249.1
I tried with round function but than I get an error saying invalid number.
Also I would like to be able to change the number of decimals I put in the test to use with different columns
For example Its 1 big table with all columns having datatype VARCHAR2(2000 char)
examples for columns:
total amount (value should have no more than 2 decimals)
48528.64
135082.54
349.1123 (not OK)
Balance (value should have no more than 2 decimals)
45428.64
1895082.11
5249.1483 (not OK)
Loan (value should have no more than 6 decimals)
100.64
88999.11654
1000.178875554 (not OK)
For each column I want to set up a seperate test that checks if the value is within the number of decimals allowed. So preferable a select statement with a where clause where I can adjust the numbers of decimals so I end up with all records having 1 or 2 decimals, or all the records that have more than 2 decimals
Upvotes: 1
Views: 897
Reputation: 142705
Invalid number
error is due to the fact that you have something that isn't a number in that column, so when you apply numeric function to it, Oracle complains. That's what you get when you store numbers as strings. Don't do that.
Anyway, here's one option which shows what you might try to do: as these are strings, calculate number of digits right of the decimal point.
SQL> select * From test;
A
--------------------
48528.64 -- OK
135082.54 -- OK
5249.1 -- OK
1.2345 -- not OK
-25.553 -- not OK
SQL> select *
2 from test
3 where length(regexp_substr(a, '\d+$')) > 2;
A
--------------------
1.2345
-25.553
SQL>
If there are several columns and you'd like to check each of them using a separate table which holds allowed number of decimals, then you could do something like this:
SQL> with
2 big (total, balance, loan) as
3 (select 48528.64 , 45428.64 , 100.64 from dual union all
4 select 135082.54 , 1895082.11 , 88999.11654 from dual union all
5 select 349.1123 , 5249.1483, 1000.178875554 from dual
6 ),
7 septest (tdec, bdec, ldec) as
8 (select 2, 2, 6 from dual)
9 select
10 b.total,
11 case when length(regexp_substr(b.total,'\d+$')) > s.tdec then 'Not OK'
12 else 'OK'
13 end total_ok,
14 --
15 b.balance,
16 case when length(regexp_substr(b.balance,'\d+$')) > s.bdec then 'Not OK'
17 else 'OK'
18 end balance_ok,
19 --
20 b.loan,
21 case when length(regexp_substr(b.loan,'\d+$')) > s.ldec then 'Not OK'
22 else 'OK'
23 end loan_ok
24 from big b cross join septest s;
TOTAL TOTAL_OK BALANCE BALANCE_OK LOAN LOAN_OK
---------- ---------- ---------- ---------- ---------- ----------
48528,64 OK 45428,64 OK 100,64 OK
135082,54 OK 1895082,11 OK 88999,1165 OK
349,1123 Not OK 5249,1483 Not OK 1000,17888 Not OK
SQL>
Lines #1 - 8 represent sample data; you already have that. Query you actually need begins at line #9.
Upvotes: 1