Reg: Performance tuning

Former Member · ‎2007 Mar 19

Hi All,

I got some Z-Objects for performance tuning. Can you please list out steps to do this.

As early as possible

Former Member · ‎2007 Mar 19

Hi venu,

First check the estimation cost for the report with basis pepole.

Check is there any nested loops or select statements are exists.

Check the indexes properly in where condition

check the fields order mention in select statement. It shuould be match with database fields order and declaration of internaltable fields also same match with data base firlds order.

Use binary search statement with read table..

Read table with proper keys.

Sort the internal tables befoe using the control break statements like at new.....

Check the internal table is empty or not before using the for all entries.

Avoid the redundancy.

Check each select statment execution time by ST05.

Hope this helps you....

Regards,

Kumar.

Former Member · ‎2007 Mar 19

Hi,

CHk this blog....

/people/rob.burbank/blog/2006/11/16/performance--what-will-kill-you-and-what-will-leave-you-with-only-a-flesh-wound

Here are some performance TIPS

LOOP AT with WHERE clause

Working with internal tables

For all entries

Tools for performance tuning

Using table buffering

LOOP AT with WHERE clause

If you use a LOOP AT statement with a WHERE clause, the table whole will be read through

and not only the entriers that satifies the WHERE clause. This can lead to performance

problems when a very large internal table is read many times with a WHERE clause.

The solution is to sort the table on the keyfields, use a READ statment to find

the first entry that satifies the key. Then you can start the loop here, and check for

changes in the keyfield to exit the loop.

Do not

loop at gi_mseg into g_mseg

where matnr = p_matnr and

werks = p_werks and

lgort = p_lgort and

sobkz = space,

endloop.

Instead use:

Sort internal table with entries from MSEG

sort gi_mseg by matnr werks lgort.

Find index of first entry that satifies the where clause

data:

l_tabix_from like sy-tabix.

read table gi_mseg with key matnr = p_matnr

werks = p_werks

lgort = p_lgort binary search

into g_mseg.

check sy-tabix > 0.

move sy-tabix to l_tabix_from.

Loop over the table from l_tabix_from, check for changes in keyfields, and

if necessary check other fields.

loop at gi_mseg into g_mseg from l_tabix_from.

if g_mseg-matnr <> p_matnr or

g_mseg-werks <> p_werks or

g_mseg-lgort <> p_lgort.

Stop loop

exit.

endif.

Check other fields

check g_mseg-sobkz = space.

..... do something ......

endloop.

For all entries

The for all entries creates a where clause, where all the entries in the driver table are combined with OR. If the number of entries in the driver table is larger than rsdb/max_blocking_factor, several similar SQL statements are executed to limit the length of the WHERE clause.

Some steps that might make FOR ALL ENTRIES more efficient:

- Removing duplicates from the the driver table

- Sorting the driver table

- If possible, convert the data in the driver table to ranges so a BETWEEN statement is used instead of and OR statement:

- FOR ALL ENTRIES IN i_tab

WHERE mykey >= i_tab-low and

mykey <= i_tab-high.

Tools for performance tuning

The runtime analysis (SE30)

SQL Trace (ST05)

Tips and Tricks tool

The performance database

Using table buffering

Using buffered tables improves the performance considerably. Note that in some cases a statement can not be used with a buffered table, so when using these staments the buffer will be bypassed. These staments are:

Select DISTINCT

ORDER BY / GROUP BY / HAVING clause

Any WHERE clasuse that contains a subquery or IS NULL expression

JOIN

A SELECT... FOR UPDATE

If you want to explicitly bypass the buffer, use the BYPASS BUFFER addition to the SELECT clause.

Kishore.