Problem with select fetching data and processing

[Former Member]

Hi Experts,

I am doing performance tuning for a program.

I am fetching data from EKET table based on EBELN , EBELP and EINDT field entries into I_EKET table.

Than I am processing the entries in I_EKET table as

SELECT EBELN EBELP EINDT

     FROM   EKET

     INTO   TABLE I_EKET

     FOR ALL ENTRIES IN I_EKPO

     WHERE EBELN EQ I_EKPO-EBELN

     AND   EBELP EQ I_EKPO-EBELP

     AND   EINDT IN S_EINDT.

     IF I_EKET[] IS NOT INITIAL.

       SORT I_EKET BY EBELN EBELP.

       DELETE ADJACENT DUPLICATES FROM I_EKET COMPARING ALL FIELDS.

is it possible to write the sort and delete adjacent duplicates inside the select query itself.

Thanks,

RG

[Former Member]

Hi Ramya G,

I think following select statement will cover your duplicates:

SELECT distinct EBELN EBELP EINDT

      FROM   EKET

      INTO   TABLE I_EKET

      FOR ALL ENTRIES IN I_EKPO

      WHERE EBELN EQ I_EKPO-EBELN

      AND   EBELP EQ I_EKPO-EBELP

      AND   EINDT IN S_EINDT.

Distinct will filter your duplicates. BUT, distinct also has a low performance.

(ABAP Performance tips - Contributor Corner [Read-only] - SCN Wiki)

AND, if you are using a HANA database, it is possible that your data will be unsorted.

()

If your data needs to be sorted and you are on a HANA database, then you will have to use the order by statement.

With a for all entry you can only use 'order by primary key'. However you do not use all the key fields in your select statement so it will not be possible to use the 'order by primary key'.

So, I think the best way to solve this, is by removing the for all entries.

I see that you are using a for all entries on EKPO. Did you do a select statement on EKPO before the select statement above? Maybe you can Join them, hereby avoiding the use of a 'for all entries' and nicely place the needed ekpo and eket data in one itab. And place a 'sort by ebeln, ebelp' and a 'group by ebeln ebelp...' in the joined select statement. This is of course only valid if you do not have a for all entries in you EKPO select statement.

[matt]

It doesn't matter whether you're using HANA database or ORACLE or any other. The order of returned records is never guaranteed to be sorted! If you want it ordered, you must specify an ORDER BY clause.

[former_member405170]

HI Ramya,

Go for Secondary index to remove any duplicate entries while fetching data.

Thanks,

Paul