What is the best way to import a huge list of transport requests(TMS / tp)?

[Former Member]

TMS Gurus,

Currently it is required to import a huge list of transports requests (approximately 4000) into several clients/systems. Please keep in mind this: there are several repeated entries in the list, and therefore it is not possible to have all the transport requests in the required order in u201Cone single import queue (buffer)u201D.

Here is a little example to explain the u201Cimport queue (buffer)u201D situation mentioned above:

(

It is required to import these transport orders in the following sequence:

D10K900001

D10K900002

D10K900001

D10K900003

D10K900004

If we add them to the buffer by the means of a u201Ctp addtobufferu201D the buffer will looks like:

D10K900001

D10K900002

D10K900003

D10K900004

)

So far we have used to different methods :

METHOD 1) TMS Standard tools (mass transports):

For this we have to manually split the main list as much as required to avoid repeated values within these sub-lists.

Example:

Sub-list one:

D10K900001

D10K900002

Sub-list two:

D10K900001

D10K900003

D10K900004

Then we add the u201CSub-list oneu201D to the u201Cimport queue (buffer)u201D and press the u201Clittle orange Truck icon (import all requests (Ctrl + F9))u201D.

After getting the u201CSub-list oneu201D successfully imported into the target system, it is required to delete the u201Cimport queue (buffer)u201D and do the same for the sub-list two.

METHOD 2) TP ADDTOBUFFER and TP IMPORT OS scripts.

For this option two scripts are required, one for adding to the buffer all the transport requests (check u201Cexample 1u201D) and other to actually get them imported in the required sequence (check u201Cexample 2u201D).

Example 1:

tp addtobuffer D10K900001 TE1 pf=/usr/sap/trans/bin/TP_DOMAIN_QE2.PFL

tp addtobuffer D10K900002 TE1 pf=/usr/sap/trans/bin/TP_DOMAIN_QE2.PFL

tp addtobuffer D10K900001 TE1 pf=/usr/sap/trans/bin/TP_DOMAIN_QE2.PFL

tp addtobuffer D10K900003 TE1 pf=/usr/sap/trans/bin/TP_DOMAIN_QE2.PFL

tp addtobuffer D10K900004 TE1 pf=/usr/sap/trans/bin/TP_DOMAIN_QE2.PFL

Example2:

tp import D10K900001 TE1 client=700 U0126 pf=/usr/sap/trans/bin/TP_DOMAIN_QE2.PFL

tp import D10K900002 TE1 client=700 U0126 pf=/usr/sap/trans/bin/TP_DOMAIN_QE2.PFL

tp import D10K900001 TE1 client=700 U0126 pf=/usr/sap/trans/bin/TP_DOMAIN_QE2.PFL

tp import D10K900003 TE1 client=700 U0126 pf=/usr/sap/trans/bin/TP_DOMAIN_QE2.PFL

tp import D10K900004 TE1 client=700 U0126 pf=/usr/sap/trans/bin/TP_DOMAIN_QE2.PFL

Edited by: Juan carlos Perez on Sep 30, 2009 2:24 AM

[Former Member · Accepted solution]

Given, I assume, that you don't have advanced third party tools such as Rev-Trac I would suggest that your method 2 is the best solution. For massive lists like this the editting facilities at the operating system level are usually much more powerful than those in the GUI based interfaces of SAP transactions.

[Former Member · Accepted solution]

Nicholas,

Thanks for your reply, you assumption is correct we don't use 3rd party tools for transports.

I agree with you, method 2 is easier to implement than method 1 because no manual splitting of the main list is required. But what about performance?

Method 2 takes longer to get all trasports requests imported. Our records shows that a maximum of 7.5 transport requests get imported in one minute. That is approximately 450 transport requestsin an hour.

Method 1 is faster, it could help us to save time, the import of the same list could be done in half the time (almost 1000 transport requests in an hour). But the time saved could be wasted while doing the manual splitting (required in case of repeated entries), the deletion and population of the trasnport queue. And as you know manual modifications can be dangerous if one single transport request is omitted or misplaced by error,

We have read the following SAP notes:

Note 1127194 - R3trans import with parallel processes

Note 1008058 - Performance of individual imports.pdf

Note 1223360 - Composite SAP Note - Performance optimization during import

We would like to have the best from these to methods:

Method 1: Fast

Method 2: Just one list.

How performance can be improved for method 2?

Is it possible the trigger parallel tp import process while using u201CMethod 2u201D?

Thanks once again for your valuable help.

[Former Member · Accepted solution]

I haven't read the notes to which you refer. Generally speaking, however, transports are usually sequenced for a particular reason. If I import, for example, the following:

SIDK900123

SIDK900124

SIDK900125

SIDK900123

it's usually because SIDK9000123 contained an object which could not be fully activated until something in 124 or 125 was imported. The probability exists in this circumstance that 124 and/or 125 may well have a dependency on something in 123 so the sequence and the repeated import must be preserved. In this situation it would be very difficult to do parallel imports without first performing a very detailed analysis of the dependencies between your subsets of transports.

I still think your method 2 is the safest. It's easiest and least error prone to set up and, given that you know that the result of importing these transports in this sequence has delivered the result you want, it's the only process which can be reliably claimed to give you the result you want in the next target system.

Good luck,

MURRAY