Finding out module from incident numbers using text parsing

[manish_malakar0316]

Hi all,

I have been given a requirement where there are lots of incident numbers and their descriptions in an excel file and I have to develop a report to show in which modules (SD,MM,FI,PP) each incident belongs to. I have attached the text file with this thread(it is actually in excel format but excel cant be attached)

<removed by moderator>

I was told by the lead technical architect ( the one who assigned this task) to use text parsing algorithm for this. Can anyone tell me how text parsing can help me to identify the modules for each incident ?

Regards,

Manish

[manish_malakar0316]

Hi all,

I have been given a requirement where there are lots of incident numbers and their descriptions in an excel file and I have to develop a report to show in which modules (SD,MM,FI,PP) each incident belongs to. I have attached the text file with this thread(it is actually in excel format but excel cant be attached)

<removed by moderator>

I was told by the lead technical architect ( the one who assigned this task) to use text parsing algorithm for this. Can anyone tell me how text parsing can help me to identify the modules for each incident ?

Regards,

Manish

[ArthurParisius]

Looking at the file, I don't think there is a way to do it. Unless your company makes some agreements on how the description of the incidents need to be made.

[SimoneMilesi]

I doubt you can do it without having a bunch of entries in a table connecting all the possible key words (like "PO" for "PO 420686992 Freelance Caroline Receveur - Button "Modify" or ELIXPEDIA for "Missing prices in ELIXPEDIA") to the reference module and even in this way you'll have a ton of incidents you cannot identify (i.e. Applications métiers pour AUDOUIN-URDANGARAY Julie ).

[manish_malakar0316]

Exactly . Even if I search strings based on a particular module, how am I supposed to find out the modules of some incidents like below which are not even in english:

INC0926497 PEDIDOS EN IDOCS

INC0927197 Manque intégration facture pour faire un retour usine

INC0927279 Probleme lors de la création de DED

INC0927299 MANQUE DE NUMERO DE FACTURE POUR RETOUR USINE

Then there is one incident with just this:

INC0927288 Jobs cancelled

How will I know which module's job got cancelled ???

So can this requirement be done at all ?

[Sandra_Rossi]

Don't post productive data that might contain sensitive data (you probably didn't make this file approved by your company for public release; they would probably refuse). Just take a few examples that you are sure are anonymized and netetiquette.

[manish_malakar0316]

ok understood sandra.rossi , but do u think that there is a solution to figure out modules based on incidents like:

INC0927288 Jobs cancelled

[manish_malakar0316]

sandra.rossi incidents like:

INC1008305 Delivery created for document blocked by workflow

INC0930545 Intégration IDOC catalogue en emergency

INC0930853 PEDIDOS ENVIADOS DE SAP A TMS, PERO NO SE RECIBE EN ESTE SISTEMA. ADJUNTO FICHERO CON LAS ENTREGAS AFECTADAS.

which translates to "ORDERS SENT FROM SAP TO TMS, BUT NOT RECEIVED IN THIS SYSTEM. ATTACHED FILE WITH THE DELIVERIES AFFECTED"

[Former Member]

Hi,

I don't see how this has got anything to do with ABAP.

To solve it you need to link strings to modules and also number patterns corresponding to number ranges (e.g. 46######## is a PO number and therefore MM) and then do a stringsearch in the description cel for all of those strings and patterns. This can be done in excel with basic formulas.

Kind regards, Rob Dielemans

[mynynachau]

Please attach a new file that is not containing sensitive information. The original file was removed.

[manish_malakar0316]

mynyna.chau I dont know which information is sensitive in the file. There are over 11,000 rows in the original file , so I cant filter out which are sensitive or non-sensitive ones. But can you tell me if there is any feasible solution for this problem or not ? I already got a couple of replies stating that it cant be done via ABAP.

[mynynachau]

I'd like to quote another community member Sandra Rossi that has already recommended a way how to do it instead:

"Don't post productive data that might contain sensitive data (you probably didn't make this file approved by your company for public release; they would probably refuse). Just take a few examples that you are sure are anonymized and netetiquette"

[NTeunckens]

I agree with Rob Dielemans : why do this in SAP / ABAP, having to code your way around the specifics on the SAP Modules and its main Objects, Number Ranges etc.

In my experience, even if you have successfully coded this in ABAP for SAP-Modules 'X-Y-Z' that you are knowledgeable of ... you guessed it ... now they want it for SAP-Modules 'Q-R-S' as well, which you might know little or nothing about ...

[matt]

Treat it as a "best effort". You're attempting to help get incidents properly categorised and using the program to help you do this. But it can't do all the work. Figure out some rules, like the ones others have identified that can categorise at least some of the incidents.

Alternatively, get a bunch of experts together, and a few thousand incidents and categorise them using human intelligence. This way they can be used as training data for an AI tool.

Ensure the technical lead understands that there will be mis-assignment, and that many incidents will not be able to be assigned at all - no matter what you try. If he doesn't comprehend why, then suggest to his boss that he's needs a new technical lead. One who is competent.