Re: SQL Indicator Arrays

[Vernon Hamberg <vhamberg@xxxxxxxxxxxxxxx>]

First off, I think I am somewhat confused about these things. I know I 
ran into some confusion a couple weeks ago when I was trying to use 
tests for NULL attributes.

And I wonder if this stuff is part of the tightening of SQL constructs 
done around 6.1 - here is a statement from V5R2 of the SQL Reference:

    "The reference :S1:S2 is a host structure reference if S1 names a 
host structure. If S1 designates a host structure, S2 must be either a 
small integer variable, or an array of small integer variables. S1 is 
the host structure and S2 is its indicator array."

That suggests we CAN use an array that is of the same size as the number 
of elements in the structure - in fact, the documentation says much 
about when the sizes don't match. But this is an SQL reference, not RPG 
reference, and I didn't find an example like this.

I found that I could not use at 7.1 an array for the indicator 
variables. I had to have a structure with the same number of subfields 
as the DS for the fields.

Here's a thought - a data structure for the record is column-based - 
maybe the indicator "array" really has to be a similar structure of 
int(5 : 0)'s that are also column-based. As far as SQL is concerned, it 
sees it as an array - a set of contiguous values of the same data type. 
But for RPG it has to be a DS.

I see that at 7.3 that the LIKEREC and EXTNAME keywords can specify 
*NULL - this makes a data structure that matches the columns - the data 
type is indicator - so this is NOT usable for SQL.

I'd like to at least see the LIKEREC or EXTNAME keywords to have an 
option for *SQLIND (or whatever) to generate this related indicator 
"array" as a data structure - is that in any RFE?

Cheers
Vern

On 11/18/2016 10:37 AM, Peter Dow wrote:
> What else is it about?  A google for "SQL null indicator array" finds 
> it.  Are you saying /you/ use it for more than just nulls? Or the 
> system does?  If the system, what else does it use it for?
>
>
> On 11/17/2016 9:11 PM, dlclark@xxxxxxxxxxxxxxxx wrote:
> > "MIDRANGE-L" <midrange-l-bounces@xxxxxxxxxxxx> wrote on 11/17/2016
> > 05:10:27 PM:
> > > What would be nicer, imho, is if we didn't have to keep track of those
> > > null indicators - let the system do it. Provide us with a function
> > > %isNull(fieldname) or something, and let it figure it out. Sooo much
> > > simpler.
> >          To all the recent responses....  Keep in mind that the SQL
> > Indicator array is about so much more than just nulls.  This is why 
> > it is
> > NOT called the SQL Null Indicator array.
> >
> > Sincerely,
> >
> > Dave Clark