RE: SQL Join - all rows from any table for employee -- MIDRANGE-L

This isn't about encapsulation, but let's look more closely at this one as an example. After development, someone decides they want to isolate a certain department (department is not in the original spec). With SQL, no problem; we can join to the PR master, selecting only those rows that match that department. With the "native I/O" solution, it's a program change, and possibly not a minor one. (Not too bad, in the shell you presented.) Then, oh by the way, they would like it in descending order by wage base. This is at least a rewrite of your program as you presented it; addition/alteration of an ORDER BY in SQL.

Dennis Lovelady
http://www.linkedin.com/in/dennislovelady
--
We don't care. We have each other, on the Internet.
-- Dave Barry

-----Original Message-----
From: midrange-l-bounces@xxxxxxxxxxxx [mailto:midrange-l-
bounces@xxxxxxxxxxxx] On Behalf Of Michael_Schutte@xxxxxxxxxxxx
Sent: Friday, June 11, 2010 9:08 AM
To: Midrange Systems Technical Discussion
Subject: Re: SQL Join - all rows from any table for employee

Who says you cant encapsulate this code? Put it in a service program
and still get what you want.

--

Michael Schutte
Admin Professional

Announcing Bob Evans Bob-B-QÂ Road Trip! For a limited time, America's
best Bob-B-QÂ tastes are all at Bob Evans! For more information, visit
www.bobevans.com/menu/seasonal.aspx

midrange-l-bounces@xxxxxxxxxxxx wrote on 06/11/2010 04:14:27 AM:

Michael,

Sometimes it just might be easier to write the program with I/O.

Yes, of course. As always, it depends on your business needs. On the

other

hand, as I see it, one of the main advantages of doing this kind of

solution

(Birgitta's SQL solution) is that you can define a particular

business
rule

just once, encapsulate in a SQL view and have it accessible for
several developers and/or programs, queries, etc., without having to
reinvent the wheel every time (by nature, I am lazy, so this kind of
solutions appeal

to

me :-) ).

Also, there is what I suspect is the main interest in this kind of
solutions: Just to see if and how can be done. In other words, just
for

the

heck of it...

Best Regards,

Luis Rodriguez
IBM Certified Systems Expert â eServer i5 iSeries

On Fri, Jun 11, 2010 at 7:49 AM, <Michael_Schutte@xxxxxxxxxxxx> wrote:

Sometimes it just might be easier to write the program with I/O.

DoW not %EoF(PRMAST);
Read PRMAST;
If %EoF(PRMAST);
Leave;
EndIf;

Chain (empno) DEDBAL;
If not %Found();
// Clear fields;
employee_contributions = 0;
employer_contributions = 0;
EndIf;
Wages = 0;
SetLL (empno) INCBAL;
Dow not %EoF(INCBAL);
Read INCBAL;
If %EoF();
Leave;
EndIf;
// sum wages
Wages += income_amount;
EndDo;
// Print employee data
EndDo;

--

Michael Schutte
Admin Professional

Announcing Bob Evans Bob-B-QÂ Road Trip! For a limited time,
America's

best

Bob-B-QÂ tastes are all at Bob Evans! For more information, visit
www.bobevans.com/menu/seasonal.aspx

"Dennis Lovelady"
<iseries@lovelady
.com>

To

Sent by: "'Midrange Systems Technical
midrange-l-bounce Discussion'"
s@xxxxxxxxxxxx <midrange-l@xxxxxxxxxxxx>

cc

06/10/2010 06:30

Subject

PM RE: SQL Join - all rows from

any

table for employee

Please respond to
Midrange Systems
Technical
Discussion
<midrange-l@midra
nge.com>

Yeah, I was pretty loose about the specifics because I don't want

to
give

the impression that I want someone else to write this for me. I

don't.

But
here we go:

DEDBAL contains quarterly deduction amounts (employer and employee
contribution), the year, the quarter, employee number and other

stuff.

INCBAL contains quarterly income amounts, Federal wage base, the
year,

the

quarter, employee and other stuff.

I need employee name, year, quarter, max(wage base), sum(income

amounts),

sum(employee contribution), sum(employer contribution) and some
other stuff.

So from these three tables, I need multiple columns of each, put

together

onto a single row by employee.

UNION might work, but I don't think UNION DISTINCT (by itself) will
do

what

I want. I could do a left outer and an exception join, but then

the

summing and other details would get complex and would need to be
mostly (but

not

completely) duplicated. Messy.

I liked the look of FULL JOIN, but apparently that's only available

some

after V5R3, which is where we are.

Likely I'll just write a program, but I wondered if there might be

a
more

standard solution.

Sorry I wasn't completely upfront before. I just don't want to be
one

of

those "here's my task - do it for me" people.

Dennis Lovelady
http://www.linkedin.com/in/dennislovelady
--
"A cynic is a man who knows the price of everything and the value

of

nothing."
-- Oscar Wilde

Do you need info from DEDBAL and INCBAL? If not, you could use

EXISTS.

SELECT * <-- desired fields from PRMAST go here FROM PRMAST

WHERE

EXISTS (SELECT 1 FROM DEDBAL WHERE DEDEMP=PREMP)
OR EXISTS (SELECT 1 FROM INCBAL WHERE INCEMP=PREMP)

If you need to pull data from DEDBAL and INCBAL, you could use a
LEFT OUTER join and then exclude those cases where an employee

has

no records in

DEDEMP

and
INCEMP.

Try this:

SELECT * <-- desired fields from PRMAST, DEDBAL, and INCBAL go
here FROM PRMAST LEFT OUTER JOIN DEDBAL ON DEDEMP=PREMP LEFT

OUTER

JOIN INCBAL ON INCEMP=PREMP WHERE DEDEMP IS NOT NULL OR INCEMP IS
NOT NULL

In this case, if an employee has records in DEDBAL but not in
INCBAL, any fields in INCBAL would be null and vice versa.

Have fun!

Richard Casey

-----Original Message-----
From: midrange-l-bounces@xxxxxxxxxxxx
[mailto:midrange-l-bounces@xxxxxxxxxxxx] On Behalf Of Dennis
Lovelady
Sent: Thursday, June 10, 2010 3:38 PM
To: 'Midrange Systems Technical Discussion'
Subject: SQL Join - all rows from any table for employee

We have three tables: DEDBAL, INCBAL, and PRMAST.

Each of these contains an employee number (DEDEMP, INCEMP, PREMP).

We

want
to produce output where an employee is represented in either
DEDBAL

and

INCBAL or both. (For the sake of argument, we can assume that
PRMAST is present for each employee).

I've tried various types of joins (LEFT OUTER, FULL, et cetera)
and have pored through the archives and GOOGLE and come up empty
on a means of doing this exactly right. (I'm convinced this is a
common issue; so my

issue

is
probably my choice of search words - usually resulting in too

many

hits).
Would some kind soul mind directing me to guidelines on how to
write the JOIN for this challenge?

Thanks!

Dennis E. Lovelady
AIM/Skype: delovelady MSN: fastcounter@xxxxxxxxxxxx
<http://www.linkedin.com/in/dennislovelady>
www.linkedin.com/in/dennislovelady -- I am at one with my duality.

--
This is the Midrange Systems Technical Discussion (MIDRANGE-L)

mailing

list
To post a message email: MIDRANGE-L@xxxxxxxxxxxx
To subscribe, unsubscribe, or change list options,
visit: http://lists.midrange.com/mailman/listinfo/midrange-l
or email: MIDRANGE-L-request@xxxxxxxxxxxx
Before posting, please take a moment to review the archives
at http://archive.midrange.com/midrange-l.

--
This is the Midrange Systems Technical Discussion (MIDRANGE-L)

mailing
list

To post a message email: MIDRANGE-L@xxxxxxxxxxxx
To subscribe, unsubscribe, or change list options,
visit: http://lists.midrange.com/mailman/listinfo/midrange-l
or email: MIDRANGE-L-request@xxxxxxxxxxxx
Before posting, please take a moment to review the archives
at http://archive.midrange.com/midrange-l.

--
This is the Midrange Systems Technical Discussion (MIDRANGE-L)

mailing
list

To post a message email: MIDRANGE-L@xxxxxxxxxxxx
To subscribe, unsubscribe, or change list options,
visit: http://lists.midrange.com/mailman/listinfo/midrange-l
or email: MIDRANGE-L-request@xxxxxxxxxxxx
Before posting, please take a moment to review the archives
at http://archive.midrange.com/midrange-l.

--
This is the Midrange Systems Technical Discussion (MIDRANGE-L)

mailing
list

To post a message email: MIDRANGE-L@xxxxxxxxxxxx
To subscribe, unsubscribe, or change list options,
visit: http://lists.midrange.com/mailman/listinfo/midrange-l
or email: MIDRANGE-L-request@xxxxxxxxxxxx
Before posting, please take a moment to review the archives
at http://archive.midrange.com/midrange-l.