When implementing an e-business integration solution, whether it is cloud, data, application-to-application (A2A) or business-to-business (B2B), there are always common activities that have to be performed. One of those activities is creating the data transformation maps. This is usually the most complicated and time consuming part of the integration process. In an attempt to reduce time and effort required to create these maps, different approaches, both manual and automated, have been tried and tested. Let’s take a look at some of these methods and which circumstances to use them:
One available approach is the process of identifying the source and target data formats, then manually mapping the rules for the transformation. Unlike the old conventional method of coding, current integration middleware solutions usually employ some form of visual drag-and-drop functionality. This is the most common, but also the most time-consuming approach. Manual mapping can be used for building new transformations, especially when no similar map is available to use as a starting point. This approach can also be used when implementing a brand new integration middleware solution, or when there aren’t any other available options. Examples of this are when new applications and trading partners are integrated into the solution or, when brand new document formats from external integrations are introduced.
Another manual approach is map reuse. This is the practice of copying and modifying an existing map to suit a new transformation requirement. The base map is the starting point for the new map and usually contains mappings that are most common to the type of transformation being performed. In other words, it is basically a “best fit” transformation map that uses the same source and target document formats. The most common utilization of map reuse is with B2B integrations that use electronic data interchange (EDI), Rosettanet, or some other standardized format that connects to another common format, like back-end application files, or even interface files to an ERP. One thing to note about map reuse is that you always need a base map to copy from. Many vendor supplied ERP interface adapters or connectors contain sample maps for reuse.
Automated or Assistive Mapping
There are many names that identify automated or assisted mapping, such as “smart mapping”, “mapping suggest”, “auto-mapping” or, “just-in-time assistance”. This is the practice of utilizing assistive tools in order to automate parts of the mapping process. It works by suggesting matches between the source and target documents based on the different suggestion algorithms available. These tools vary from vendor to vendor and utilize different methods such as:
- Property matching: Using names, data types and other properties from the source and target formats to suggest a mapping instruction
- Dictionary-driven: Referencing user-defined dictionaries that relate possible matches based on key-value pair look-ups
- Metadata tagging: Searching user-defined tags on the data nodes (fields, elements, etc.) of the source and target documents
- Mapping patterns: Using mapping histories from previously created maps
Some solution providers, like EXTOL, additionally provide cloud resources to create a community-based mapping pattern repository. One positive aspect of the mapping assistive approach is, in most cases, it can be utilized anytime for any type of integration. If you are implementing a new e-business integration solution, you can immediately take advantage of the property method to create your first map. You can also take a little bit of time up front and create a dictionary and/or tag your metadata prior to creating your first map. If your solution provider provides a cloud-based mapping pattern repository, you can immediately create a map based on some of the best practices from other mapping histories. If you have already created a few maps using the map reuse approach, you can use the mapping patterns from the modified base maps to create a more robust map that will most likely need less modifying.
Keep in mind, there isn’t a right or wrong mapping approach. All of the manual and automated approaches listed above serve their intended purpose when they are needed.