extracting data from faltfiles

[Former Member]

Hi BWers,

When extracting data from faltfiles, how can we know that is a new record or modified record.How to model these in BW.

Sekhar

[Former Member · Accepted solution]

Hi Chandra,

By Using Time Characteristics you can identify the records are new or modified.

as per my knowledge, Change log table will have the complete history of data, there you can identify new or modify.

Cube will change log table, it will maintain the history of data.

Hope it helps.

Regards,

varun

Message was edited by:

Varun Jain

[Former Member · Accepted solution]

Hi varun,

This means when ever we extract data from FF we should first load it in ODS right.

[former_member185071 · Accepted solution]

Hi Chandra Sekhar

Yes via ODs is better and its in safer manner to maintain the updated records...

Regards

R M K

• Assining points is the only way of saying thanx in SDN ***

[Former Member · Accepted solution]

Hi,

Varun told that 'Time Characteristics you can identify the records are new or modified' .How it is identified can you pls give me a scenario.

Regards,

Sekhar

[Former Member · Accepted solution]

Hi,

I hope, he is talking with reference to R/3 or external data base which can maintain timestamp of last modification. And I hope, this is not possible with FF. So definately you need ODS , if you are sure that there is chance of getting duplicate records from flat file.

With rgds,

Anil Kumar Sharma .P

[Former Member · Accepted solution]

Hi sharma,

Still I'm not clear how deltas are handled in FF.Can you send me some docs regarding this.

[Former Member · Accepted solution]

Hi,

Deltas are captured in ODS , not is FF.

To capture the Deltas , we need to upload to ODS from FF first. So the system compares the Active table records of ODS with new table (newly uploded ) records of ODS, then enter the Deltas in Change log table of ODS which is used in delta load to further data targets.

http://help.sap.com/saphelp_bw31/helpdata/en/10/54853b175d214ee10000000a11402f/frameset.htm

with rgds,

Anil Kumar Sharma .P

[Former Member · Accepted solution]

Hi

Generally ODS delta updation takes place Only using 0recordmode

The field 0RECORDMODE determines whether the records are added to or overwritten. It determines how a record is updated in the delta process: N' for New data, blank character signifies an after image, ‘X’ a before image, ‘D’ deletes the record and ‘R’ means a reverse image.

When you are loading flat files you have to select a suitable delta process from the transfer structure maintenance, this ensures that you use the correct type of update.

there are 7 types updation in a record mode

' ' After image

X Before image

A Add

D Delete (Key only)

N New image

R Reverse image

Y Update image (Values for update)

In Record mode

there is two types of Updation 1. Additive mode and Overwrite

If it is a Overwrite mode latest value will be saved in a target

If it is a Addition Mode the value will be Added to the particular field

This can be viewd in change log table.

[Former Member · Accepted solution]

Thanks kalpana,

I have an another question ,How duplicate records are handled

[Former Member · Accepted solution]

Hi

Mainly duplication occurs maily because of some calculation fault in update rule aor rotines, if it mismatch with data in same records the recods will splited into two thats why there is duplication takes place most of the time.....this can be identified by using cube>rclk>manage-->and find the no of records added and tranferred i transferred is equal to added no prbs. if added is greater than transferred and this can be also find(exactly where this could occur) by debugging....

[Former Member]

I load an infocube using flat files in a Datasource that have the Characteristic RECORDMODE and this have the possibility to mark the record as "For deletion" and the record new come in the same file and replace the last value