Re: Still problems with SQLCLI

Hi Joep, Thankyou for taking the time to take a look. This is very much
appreciated.

The reason I use DS_COLMETA.LENGTH.INTEGER for the SQLBindCol() call, is
that for numeric type columns, this field stores both the size, and scale
(length and decimals) of the field to be bound. The value is calculated as
SIZE*256+SCALE, or, for example, a currency type field of format 10
digits, 2 decimals, should be passed as 2562.

( I do not calculate this value, I just take the value given to me, from
when I check the format of the returned select statement).

The field length for a 10,3 field, is, of course, just 10 bytes (for zoned
numeric fields). This is why I use DS_COLMETA.SIZE to increment the
pointer/record size.


Cheers,




Tony Weston
Senior Analyst/Programmer - IT Systems


Please scan all attachments for viruses.




J.Beckeringh@xxxxxxxxxxxxxxxxxxxxxxxxxx
Sent by: rpg400-l-bounces@xxxxxxxxxxxx
19/06/2008 14:25
Please respond to RPG programming on the AS400 / iSeries

To: RPG programming on the AS400 / iSeries
<rpg400-l@xxxxxxxxxxxx>
cc:
Subject: Re: Still problems with SQLCLI


Hello Tony,

I had a go at it, buf of course my time is limited.

What I noted is:

W_rowstsarray elements 95 en 96 are indeed screwed up and the watch in the

debugger does not notice it

Something fishy is going on with SQLBindCol: You pass a pointer
W_REC_BUF_PT and a length DS_COLMETA.LENGTH.INTEGER, buf after the call
you increase W_REC_BUF_PT with DS_COLMETA.SIZE. I changed that to
DS_COLMETA.LENGTH.INTEGER. W_rowstsarray stays intact, but then the
program starts to loop and when I finish it, the joblog says: 'Processing
of the SQL statement ended. Reason code -725366539.' That is not healthy.

It has to be something with a pointer, of course.

Joep Beckeringh


rpg400-l-bounces@xxxxxxxxxxxx wrote on 19-06-2008 15:03:21:

[image removed]

Still problems with SQLCLI

Tony.Weston

to:

rpg400-l

19-06-2008 15:05

Sent by:

rpg400-l-bounces@xxxxxxxxxxxx

Please respond to RPG programming on the AS400 / iSeries

Hi,

I am still having problems with my SQLCLI program.... No replies for my
last post.... Surely some technical guru here should know what I'm doing

wrong!...

The program in question is here: http://totspics.com/sqlcli.txt

To see the problem:

**Compile, Run in Debug with a breakpoint on the first
SQLFetchScroll()

**When the program stops at the breakpoint, check the value of the
array w_RowStsArray . This will have 100 elements, all of value '11111'
(as initialised)

**Step over the SQLFetchScroll, and check the value of array
w_RowStsArray again. This time, elements 95, and 96 will be corrupted.


w_RowStsArray is not used anywhere in the program, yet is been

corrupted.

As it stands, In the actual program, (which this is based on) I have to

fetch records from the remote server 1 record at a time, which is very
slow, and the '.net people' in the office are begging to take the 'P'!
Help me save face!....Anyone!

Many thanks in advance, to anyone able to offer any suggestions.





Tony Weston
Senior Analyst/Programmer - IT Systems