Re: Handling decimal data errors at the field level

Thomas Garvey wrote:

... the IBM manuals indicate that when a record is read directly into a
data structure, the null indicators are not set. I think that's fixed in
v5r4 but I'm writing for v5r3 at least.

Even in V5R2, when a record was read directly into a data structure, the
null indicators would be set for the data structure subfields.

This was a late-breaking change in V5R2, probably too late to get into the V5R2 manual.

But here's what it says in the V5R3 reference on the "User Controlled Support for Null-Capable Fields and Key Fields" page,
http://publib.boulder.ibm.com/infocenter/iseries/v5r3/topic/books_web/c0925085203.htm

"When a record format or file with null-capable fields is used on an input operation (READ, READP, READE, READPE, CHAIN) and a data structure is coded in the result field, the values of %NULLIND for null-capable data structure subfields will be changed by the operation. The values of %NULLIND will not be set for the input fields for the file, unless the input fields happen to be the subfields used in the input operation."