The migration workbench has 3 main steps to be processed:
- Import data
- Convert data
- Import the data
- Import data
(The legacy system or systems delivery data to be migrated, in any (file-based) form. The date is imported and stored in the migration workbench as a "read file" (unconverted); - Convert data
(From a source format to a target format, which can involve translation tables, determination logic for missing information and checking the source data for validity); - Import the data
(Apply the data to the SAP system, in one of a viariety of manners, as Idoc, BDC recording, direct input report or Bapi call).
A pre-requisite to working an LSMW step plan is the customizing in SAP - this should be completed to avoid having to rework your migration steps.
Import data
The data to be imported is in a format which is defined by the "supplying party", thus it can be one of many ways. Detail like the sequence of columns in files, field separators (or predefined field lengths), or complex file content where the actual line content is controlled by a record-identifier - it's all supported.
The workbench allows you to describe how a file or several files are to be interpreted, before they are processed into a read file. This is not actually focusing on field content, but focusing on labeling the fields (giving them actual names).
Let's assume a simple conversion of header data and item data. If the information is delivered in 2 files, the relation between the 2 datasets can be defined in the workbench as part of the import.
Convert data
To convert the data the source and target structures are mapped together in a tree-like setup. Abap development is required here, unless the conversion is very straight-forward. Converting data can involve translations, which can be maintained separately within the same project/subproject object. The end result of "Convert data" is again a file or set of files, but this time it should be prepared to be processed by one of SAP's data-take-on ways (Idoc, Bapi, ... ).
Import the data
Once all preperations are done, the actual import can be started. This will use the converted data as import file for whichever import method was chosen. As the import methods can all perform their own validity checking, the data integrity is (and remains) secured. Invalid data will simply not be accepted!