Infosys’ blog on industry solutions, trends, business process transformation and global implementation in Oracle.

« Emerging Trend in SSHR- PART3 | Main | Oracle Data Visualization Cloud Service »

Transaction matching of around two million records in under 5 minutes in ARCS

Oracle Account Reconciliation Cloud Service (ARCS) with Transaction Matching is a cloud based reconciliation platform with pre-built configurations and adherence to industry best practices; a recommended solution to cater to your reconciliation and matching needs.

Transaction Matching is a module within ARCS which inherits the features that facilitate preparation and review of reconciliations.

  • Additionally, Transaction matching adds efficient automation of the detailed comparison of transactions in two or more data sources
  • The calculation engine allows for intuitive "netting" of transactions within each data source to provide output which is easy to work with
  •  Flexibility in the setup and timing of the process allows to minimize the effort during "crunch time" and reduce risk

 

Transaction Matching Use Cases

Typical Transaction Matching Use Cases are shown below.

 

Use Cases.jpg

Often clients need to match more than million records between two source systems with complex match set rules. We have seen clients spending hours to try to manually match them in excel or use some solutions like Access database, Oracle tables etc. which can be very time consuming and have data quality issues. We will share our experience and some insights on how we successfully loaded and matched two source files with around 2 million records in less than 5 minutes using Transaction matching feature of ARCS for one of our e-commerce client.

Idea Inception

Client wanted to match up to 2 million records from their point of sale system (POS) and the details obtained from Merchant transaction system. They were using access data base for this activity which was giving them results in hours and they reached out to Infosys with this requirement to help them streamline this time-consuming and frustrating process.

 

Solution and Approach

Source Files.

1. Point of Sale transaction file.

    The POS file had 9 columns and the file provided was in txt format (a pdf report converted into a text file). Below is the snapshot of the same.

POS.jpg

2. Merchant system transaction file

           The Merchant system transaction file had 21 columns and the file was in csv format. Below is the snapshot of the file.

Merchant.jpg

Matching rules

Client wanted the matching rules to be based on the condition that the card number and the amount from POS transaction file matches against the cardholder number and amount from the Merchant transaction file with the stipulation of many to one transaction match where many transactions from Point of Sale system matches with single batch (grouped by amount) transaction from Merchant system file.

 

Initial Challenges

The initial challenges with this requirement are below

1. Size of File.

    The size of the files provided were huge as there were 9 and 21 columns respectively and both the files had around 2 million records resulting in file sizes of > 1 GB per file. This much large a file is difficult to read and edit by any text editor.

2. Formatting

    Another bigger challenge was formatting the given files as per ARCS transaction matching needs. The files provided were in text format and to read and format them given their file size was a tough nut to crack.

 

Infosys Solution

We took this challenge and delivered as promised. The biggest challenge was to import the file containing about 2 million transactions into the ARCS Transaction matching from both the system and match them automatically in quick time. Other tools and custom solutions were taking hours for this process. Importing 2 million records in a csv file is a huge input for any system to ingest. It would typically take anywhere between 15-30 minutes just to import one file into a system. We had another challenge in formatting the files because the file we received was a .pdf file converted into text format and we needed them to be converted into .csv to be accepted by ARCS Transaction Matching. We used Oracle ARCS TM, formatting tools, text editors and Oracle provided EPM Automate utility to format the files, automatically ingest and auto-match the files from two transactional systems.

 

The EPM Automate Utility enables Service Administrators to remotely perform tasks within Oracle Enterprise Performance Management Cloud instances and automate many repeatable tasks like import and export metadata, data, artifact and application snapshots, templates, and Data Management mappings.

 

Tips and Lessons Learnt

With the above requirement's implementation, we have learned a few lessons and below are some tips when implementing similar type of solution.

  • ARCS TM also accepts .zip format input files, hence compress the files into .zip format so that they are smaller in size plus quick and easy to upload on the ARCS cloud.
  • Powerful text editors like Notepad++ or Textpad when formatting the files, could be used.
  • Create custom attributes which can be used in matching rules for faster auto-matching of transactions.
  • If possible, try to get the export from the transactionsystems in .csv format to reduce conversion times.

Performance Metrics

Below are our performance metrics while implementing client's requirement of matching around 2 million records using Oracle ARCS Transaction Matching.

 

Import POS million records - 27 seconds

Import Merchant million records - 61 seconds

Run Auto Match - 53 seconds

 

Complete Process - 2 minutes 21 seconds (Less than half of 5 minutes)

 

Result?

 

Happy client and Happy us.

 

We deliver!!!! - Please visit our company website to know more about our Account Reconciliation and Transaction matching solutions.



Post a comment

(If you haven't left a comment here before, you may need to be approved by the site owner before your comment will appear. Until then, it won't appear on the entry. Thanks for waiting.)

Please key in the two words you see in the box to validate your identity as an authentic user and reduce spam.

Subscribe to this blog's feed

Follow us on

Blogger Profiles