Skip to Content

                                                           Basics of PIPE Monitoring Part – II

Introduction

In last blog we discussed about how to monitor transactions and correct them individually. However, error at bulk volume of data will make it difficult and time consuming to correct data in bulk. In this second part of Basics of PIPE monitoring we will discuss handling data at bulk level.

 

Scenario 1

When retail solution is freshly implemented and you are in the process of stabilizing production system by minimizing errors in the PIPE for small codes and articles. In this case it is required to keenly monitor PIPE and report all the missing articles and small codes to master data team or business. This can be done by using SAP given program. We will explore this in case of missing or unknown article.

 

Step1

Go to transaction SE38 and execute the program /POSDW/DISPLAY_MESSAGELOG.

image

Step 2

In the screen you can choose error messages based on store and/or posting date. Message no.5 will give all the article errors.

image

This program lists out all the materials for which PIPE gives error message. It will also display all the relevant information like store number, posting date, task the message corresponds for, date, time etc. This list can be managed further by exporting to excel. This will be a quick way to collect information of large number of materials. Other errors can also be extracted by using different message number which can be seen in PIPE monitoring work bench like message number 9 for discount type error.

image

Since numbers of materials are huge it is difficult to change these materials one by one at article identifier level. As described in the Basics of PIPE Monitoring Part – I. Hence, it is easier to delete these materials from PIPE and reprocess them again rather then correcting one by one. However, before to delete we need to cancel out transaction which have been processed successfully for this store and posting date combination and available for reporting in BI. The cancellation procedure is described under scenario 2.

 

Scenario 2

A set of data for store and posting date combination is incorrect in PIPE it self and same has been processed to BI. In order to rectify the problem we need to reprocess the correct data. Before, we also need to nullify the wrong data which resides in to BI reporting system. This is achieved by sending cancellation to BI which sends negative image of previously sent data resulting numeric value as zero for the data set in the BI reporting.

 

Step 1

The data which gets processes successfully from PIPE show status ‘Completed’ with task status 4. The available task status are shown below in the transaction /N/POSDW/TSCH.

image

 

image

We need to change task status from ‘Completed’ to ‘Ready to be cancelled’ for desired set of data.

Step 2

image

The changed task status can be observed in monitoring screen as shown below.  The changes made above are applied to the data set. All the transactions are now ready to be cancelled.

image

Step 3

To perform cancellation we need to dispatch this data to BI by using transaction /N/POSDW/PDIS. The dispatch process again can be controlled by desired tasks. Here Task T001 is for sending aggregated data to BI.

image

After executing below dispatcher the data will be available in BW delta queue which can be picked up by Info package. This nullifies previous data existed in the BI reporting system. Nullification is required to avoid duplicate data in the BI system as this data again has to be processed.

Once data successfully reaches to BI reporting layer and confirms that reports are displaying key figures value zero. We can now proceed with deletion of this data from PIPE and to reprocess it. The deletion can be performed using transaction N/POSDW/DELE.

image

Note:

This document intends to share author’s PIPE monitoring experience to help beginners in monitoring and error corrections in PIPE. The more efficient ways of performing these tasks are welcome to be shared.  

To report this post you need to login first.

2 Comments

You must be Logged on to comment or reply to a post.

Leave a Reply