Sei sulla pagina 1di 2

case1 when executed on 16th & 17th april without entering any data except the pe

riod date passed automatically through the prog


got a solution for this if i use primary keys then the data picking time is redu
cing but previously the data is picked using the AUFNR which is not a key field.
now the task is to find out how to pick the data by passing RUECK and RMZHL fiel
ds
if we have viafvc in that we have aufpl and rueck so that i can contact the afru
but through which table should i contact the viafvc. there should be a table wi
th aufpl and aufnr and data too in it.
if not viqmel is initial.
select aufnr aufpl from afko into table lt_afko
where aufnr = im_viqmel-aufnr.
if not afko is initial.
select aufpl rueck rmzhl from viafvc into table lt_viafvc
for all entries in lt_afko
where aufpl = lt_afko-aufpl
and rmzhl ne space.
endif.
loop at lt_viafvc into lw_viafvc.
and send the data of aufnr from afko and rueck from viafvc to lt_rueck. ok
then use the lt_rueck to get the data from afru for fast picking
endloop.
endif.
Question raised by PM
Users are complaining that the Tcode ZIW38 is very slow in Q49.
It works fine in Prod. Can you investigate why it is so ?

Start by comparing P and Q versions.
Investigated and Answered by Sreekar
Hello Madhulika
Yes, it is very slow in Q49. I had First checked with the versions in X,Q,P 49's
to make sure the program is same in all three servers.
Through Debugging and SQL trace i had found that AFRU table is taking more time
because there no key fields passed in where condition.
I had also checked the difference in SE11 to make sure how the records are displ
ayed when only aufnr is passed and when only rueck is passed.
The execution time is reduced when the data is passed through rueck.
Solution to achieve this through program lines could be as, We have to use these
two tables which have the fields: AUFNR, AUFPL in AFKO and AUFPL, RUECK in VIAF
VC.
Already VIQMEL-AUFNR is in where condition in select * AUFR statement, VIQMEL-AU
FNR can be used to link the AFKO table and the AFKO to link VIAFVC. After achiev
ing the data in corresponding ITAB's we need to loop the VIAFVC and read the AFK
O and move the fields(AUFNR, RUECK) to a new LT_RUECK itab and with this itab th
e records can be displayed much more fast.
Hello
But, why do you think there is no problem with the same code in P49?
Hello madhulika(not yet replied)
Server performance or data might be more in Q49 when compared to P49
p49 might be upgraded one when compared to q49
Performance issue might be from basis side too
[Nagireddy, Ramu]
It can because of different reasons:
1. Due to the more volume of records in QA than prod. It should not be the case
because QA will be sized equivalent to prod or less than that usually. Im not sur
e how the QA box in your organization is sized. Check the volume of records in p
rod and QA and should give you the answer.

2. If both the systems have the same volume of records, then you should look if
any index is being used during the data selection from DB. If yes there can be c
hances of wrong database index being picked up in production during the data sel
ection. You need to trace the SQL in QA and prod and can find out the difference
.

3 Finally if point 2 is fine, work with basis and see if there is any issue on t
he prod servers.

Apparently you can add an additional DB index on this table with fields used in
your select query. Analyze these options first before replying to their questi
on.
Hello madhulika
It can because of different reasons:
1. Due to the more volume of records in QA than prod. It should not be the case
because QA will be sized equivalent to prod or less than that usually. Im not sur
e how the QA box in our organization is sized.
2. In both systems versions are same that means the program lines are same. So,
I think we have to check with basis too for QA

Potrebbero piacerti anche