INSERT Performance

[Former Member]

Hello,

What is the most efficient do several INSERT

that is to say make an insert for each line or

using

INSERT <TABLE> FROM <INTERNAL_TABLE>

I have to load a specific table, i use the INSERT from INTERNAL TABLE

but the program is very very long.

When i start to update the table it is always empty.

Thanks for your help

Christophe

[Former Member]

The array insert (from an internal table) will be the quickest way.

Rob

[Former Member]

How many records will you be inserting?

Rob

[Former Member]

BETWEEN 100 000 and 300 000

i think in developpement there is no data and in integration there is less than a thousand of records.

But the program is still running for more than two hours ....

[Former Member]

Have you gone to SM50 to see what it's actually doing?

Rob

[Former Member]

YES AND I SEE NO SQL EVENTS, no DUMP in ST22 as in SM21

Christophe

[Former Member]

So the problem is likely elsewhere. If the program is not too long, would you post it?

Rob

[Former Member]

1. loop at itab.

INSERT INTO dtable VALUES itab.

clear itab.

endloop.

or

2. INSERT dtab FROM itab[].

[Former Member]

i load a table by using INSERT dtab FROM itab[].

and the response time is very long .....

I will try to INSERT line by line by i think that using an internal table seems to be the faster solution.

If anyone has a better solution, you're welcome.

My specific table has the same key than KONV and no duplicate are possible.

using a MODIFY can be a solution but i think that it will be slower than INSERT

....

[Former Member]

No - that will take longer. Try sorting the internal table by the primary key before doing the array insert.

Again - how many records?

Rob

[Former Member]

The internal table is sorted of course and it's structure is the same than target table.

[Former Member]

Problem is solved

Thank for your help