Focal Point
Problem with procedure from Data extraction from SAP to FOCUS

This topic can be found at:
https://forums.informationbuilders.com/eve/forums/a/tpc/f/7971057331/m/2511093172

February 13, 2008, 06:01 AM
VHayward
Problem with procedure from Data extraction from SAP to FOCUS
I have an ETL procedure which need to copy about 8,000,000 records from a SAP Table into a FOCUS table with Data Migrator. Now the whole proces is taking already two hours and it looks like it hung.

When I divide the whole into max amount of records of 2,500,000 each it will take 3 minutes. Which makes a total of 3 * 3 = 9 minutes in total.

How could this be, that as a whole it won't function. But if I divide it up to 2,500,000 records max. it takes 3 minutes per set?


WebFocus 7.6.2
Microsoft Windows Server 2003
February 13, 2008, 09:56 AM
TexasStingray
There are several reason that could possible contribute to this. Here are 2 that I could think of

The data is not in the same order that the focus database keys are in.

The focus database has reached its limit.

Hope this helps




Scott

February 13, 2008, 10:59 AM
Frans
Maybe try to use XFOCUS instead of FOCUS?

Frans


Test: WF 8.2
Prod: WF 8.2
DB: Progress, REST, IBM UniVerse/UniData, SQLServer, MySQL, PostgreSQL, Oracle, Greenplum, Athena.
February 19, 2008, 06:14 AM
VHayward
Well the order seems not the problem nor does it has reached it's limit. Also the use of XFOCUS isn't the solution.

So what could it possibly be? But the funny part is that when I max it to 2 million records it takes 2 minutes. Then 3 million takes 3 minutes. But more then 3 million, it hits the fan.


WebFocus 7.6.2
Microsoft Windows Server 2003
February 19, 2008, 06:34 AM
hammo1j
It could be that the optimizer of the RDBMS you are running SAP on has decided a different plan at the threshhold you point out. (Oracle optimizer is very poor from experience).

This assumes that the extract involves a number of joins and selections and is not just a straight table copy.


Server: WF 7.6.2 ( BID/Rcaster) Platform: W2003Server/IIS6/Tomcat/SQL Server repository Adapters: SQL Server 2000/Oracle 9.2
Desktop: Dev Studio 765/XP/Office 2003 Applications: IFS/Jobscope/Maximo
February 19, 2008, 06:52 AM
GamP
Vincent,

Have you been able to find out if the problem lies with retrieving the data or with loading it into a focus db?
In other words split the job in two parts. The first part retrieves the data and stores it in a flat file, the second load the file into your focus db.
That way you can try to find out what the repsonse times are per part of the process and also what the filesizes are that you're dealing with.


GamP
- Using AS 8.2.01 on Windows 10 - IE11.
in Focus since 1988
February 19, 2008, 09:27 AM
GinnyJakes
Do you have indexes in your FOCUS master? If yes, that can considerably slow down the load. There are techniques to mitigate this process, one being that you do your CREATE FILE with the master with the indexes, do the load with a master without indexes and an S0 segtype, then do a REBUILD INDEX.

But first you need to answer for yourself the question that GamP is asking, i.e where is the failure.


Ginny
---------------------------------
Prod: WF 7.7.01 Dev: WF 7.6.9-11
Admin, MRE,self-service; adapters: Teradata, DB2, Oracle, SQL Server, Essbase, ESRI, FlexEnable, Google