0000{ is converted to null instead of 0 when field type has PIC S9(5) in US_ASCII file

[Loganhex2021]

Describe the bug

In Cloud Migration Project, Cobrix library is making null for the field when we get a value like '0000{' to the type PIC s9(5) in copybook for the US_ASCII file.
In the existing on-prem system it was making 0 and not giving null for the same value.
Could you please help us to solve this

Sample US_ASCII Data:
P0000{
T12340
P89870

CopyBook:
01 Employee.
05 EmployeeType PIC X(1).
05 Salary PIC S9(5).

Expected Output:
EmployeeType Salary
P 0
T 12340
P 89870

[Loganhex2021]

@yruslan - Do you have any comment for this question ?

[yruslan]

Looks like a good catch. Will take a look

[yruslan]

Hi, I've tried reproducing the bug using a test. And it worked for me:
https://github.com/AbsaOSS/cobrix/blob/bugfix/459-parsing-ofnumerics/spark-cobol/src/test/scala/za/co/absa/cobrix/spark/cobol/source/regression/Test17NumericConversions.scala

What options do you pass to the reader (spark.read...)?

[yruslan]

Sorry, I see that you have the overpunched sign, will check that as well :)

[yruslan]

I confirm the error - sign overpunching works only for EBCDIC. We will add the support to ASCII as well shortly

[Loganhex2021]

Thanks @yruslan