Case Studies

MELT-V Data Conversion Tool

The ETL data conversion tool has been successfully used to convert highly sensitive data such as payroll and HR information, where there is no tolerance for errors. The tool was completely successful and significantly reduced the effort over manually producing conversion and validation scripts.   The importance of using the METL-V tool is that the data conversion process can be rigorously performed many times to assure data integrity before the final "go-live" conversion.  The tool reduces the time needed to create the conversion and validation scripts, allowing the conversion to be performed tens of times more than in a typical project.  As a result, the users can perform data cleansing and testing throughout the entire development phase of a project.  Essentially, the final data conversion for "go-live" becomes almost a non-event. 

Case Study #1: 

The first project was a PeopleSoft-to-Fidelity data conversion for an HR/Payroll system.  The application development took 6 months.  There were 30 tables, over 23,000 rows, and 233 data validation checks.  The source and target databases were Oracle.  The final "go-live' data conversion took only 13 minutes.  The importance of this tool on this project was that the target system was being developed over 6 months, and we were able to quickly change the output as requested by the developers, and test the conversion in a matter of minutes.  In fact, we performed well over 100 test conversions over the 6 months of development.

Case Study #2:

The second project was for a PeopleSoft application designed to store historical data about training for employees. The application was developed over a 4 month period.   The source tables were in Access and SQL*Server.  The target was SQL*Server.  There were only 7 source tables, but over 24,000 rows, and a very large number of data validation checks (383 of them).  The final "go-live" data conversion took only 22 seconds.   Again, the importance of using the tool on this project was that we could validate the data many times before the "go-live" conversion was performed, which assured a very high level of confidence in the data.

Case Study #3: 

The third project was small in size, but large in impact.  The conversion involved moving the data from Excel spreadsheets to SQL*Server tables.  While there were only 2 spreadsheets, and only 1,300 records, there was zero tolerance for mistakes, since the data was for annual salary bonus calculations.  Despite the source record store only containing 2 spreadsheets, there was a large number of data validation checks per spreadsheet (78 in all).  A serious constraint on the project was that the entire project could only take 2 weeks.  We were able to perform the conversion about 10 times during the 2 weeks, allowing the users to perform data cleansing and to identify data integrity issues in their spreadsheets.  The final "go-live" data conversion took only 20 seconds.