Following stages are generally followed in any implementation project.
1. Identifying the data to be migrated.
2. Determining the timing of the data migration.
3. Finalizing data templates.
4. Finalizing the method of data migration.
5. Freezing the design of program.
6. Deciding on migration related setups.
7. Post migration validation
Below diagram is an overview of a typical data migration process:
- Identifying the data to be migrated: The very first step of data migration is to identify different type of that data will be part of migration. This includes mainly two types of data. One is transaction data and another is master data. Example of transaction data is Purchase Orders, Sales Order, Invoice etc. Items, Item costs, Price Lists, Customers, and Suppliers etc are part of master data.
- Determining the timing of the data migration: As far as timing of data load is concerned each of the above data follows different time lines. Normally data loading starts from CRP (Conference Room Pilots) to make it perfect for final Production. UAT is conducted in an environment similar to the production environment so data migration in it is very important. The success of data migration is very much dependent on the results of UAT.
- Finalizing data templates:This is one of the important and initial steps of the design of data migration program. The template depends on data receiving from the legacy system and what all the information we need to upload into the ERP system.
- Finalizing the method of data migration: This is second step toward the design of the data migration program. In Oracle there are different ways of migration. One of the simplest methods is to use data loader. In this we directly insert the data into base tables through front-end. But this very uncommon for data migration because of following reasons:
• Only suitable for less amount of data.
• Not a timely solution.
• Regular monitoring is required.
Other normally followed methods are to use either Oracle Open Interfaces or oracle public APIs.
- Freezing the design of program: If it is planned to use open interface program or an Oracle public API, in both the cases the first step would be to write a program for SQL *Loader (A Shell Script). This loader program will load the data in staging table. The design of staging table depends on the data template. It should also include column for processing flag and for error messages. For Open Interface program data should be populated in the Oracle seeded interface tables. A custom program should be used for this purpose. This program makes sure that only validated data is going into the system. Once the data is populated into interface tables, Open Interface Program is used to upload the data into ERP system. If Oracle Public API is going to be used then there is no need to populate interface tables. A custom program should be used to call the Oracle API which will insert the validated data into base tables.
- Deciding on migration related setups: Before loading data into the system it should be insured that all the setups are there in place including setting of different parameters and profile options. Data can’t be loaded -successfully without proper set-ups resulting into poor quality data loading.
- Post migration validation: After the data is loaded into the system extracts should be collected from staging table and base tables to find the % success and to take decision for failed data.
