Start routine vs Update routine

[Former Member]

Hi to all,

pls can anybody tell me

what the difference b/w start routine(transfer rules level and update rules level) and update routine?

Pls give me one real time example as u faced in the prj

Thanks in Adv,

MSM.

[Former Member]

Hi MSM,

Start Routine process all data in the package before execution of single routine in Transfer or Update. For this reason in necessary field for result be present in Package (for example if you need to derive a particular field starting from one table you must have this field in Comm Structure also if it is not fed from Transfer). It is better than single routine from a performance point of view.

Update Routines are routines for a single field and depending from the whole combination of all other Key Fields: when you define a routine for Key Figure 1 in a cube you must define also all rules for Characteristics related to this KF, the same thing for KF2 and so on. This is a typical situation with purchasing when you define a Swuss Cheese Structure due to the definition of the single KF for every single particular Date (Order Creation - Delivery - ...).

Ciao.

Riccardo.

[Former Member]

http://help.sap.com/saphelp_nw04/helpdata/en/21/894eeee0b911d4b2d90050da4c74dc/frameset.htm

<u>Start Routine</u> - this routine is run exactly once for a data package at the beginning of the update program. The start routine is run for each data package at the start of the update. The start routine has only these parameters: MONITOR, MONITOR_RECNO, DATA_PACKAGE, RECORD_ALL, SOURCE_SYSTEM and ABORT. DATA_PACKAGE refers to the entire data package. The start routine does not have a return value. It is used to perform preliminary calculations and store these in a global data structure or in a table

<u>Update Routine</u>- This routine is run for each record by record at the update rule level

[Former Member]

HI MSM,

Start routine is used when you want to change the content if the datapackage as a whole.

The data comes in data packages. Now say example, i have one field in the datapackage record say company code and two values for this company code are coming in say comp1 and comp2. Now suppose i dont want to load data for comp2, i will delete this record in start routine. ( delete datapackage statement will do this )You cannot achieve this in update routine.

On the other hand, in the update routine, you can work on a specific field of the data package. Now suppose you have a quantity field. Depending on some other values in one table T, you want to decide the value of this field. So in the update routine, you will fetch the data from the table T, and do some calculation for this particular field, update it and proceed!!

Sriram

[Former Member]

1) A BI question please in a BI forum

2) Start routine is a routine that runs first when you start the loading (use it e.g. when you need to get the filter for the data oad dynamically). When the data is requested and handed to BI (most likely the PSA), the transfer routines enter the stage. Those routines are source-system specific.

Afterwards the update routine (writing data now from PSA to the data target) kicks in, the update routine being source-system independent.