Migrator » More Features
Migrator - Repository Check |
|
How does automation work? |
|
| In below example a project was set up for data migration from PeopleSoft v7.51 (Oracle) to v8.10 (MS SQL Server). Since the project has all database connection details configured, the "Repository Check" can directly connect to each of the source and target database for retrieval of table structures. For example purposes we assume that the table "PS_COST_UNIT" structure has not been retrieved from our source database, so a filter was set to that specific name. By default the wizard only has Step 1 and Step 2 checked since both source and target structural information needs to be available before the wizard may also create automatic mappings (Step 4). However, at this point also Step 3 could be checked as well as to create a program entry. So currently it'll only retrieve table information and it's fields (columns). |
|
 |
|
| In case of a PeopleSoft environment, we'd set the filter name to just "PS_", retrieve the table names and click "Check" to let the Migrator retrieve all matching tables in one process, which will take some while. This would be done first with the source and secondly for the target database (if available). If the use of the Migrator would be limited to export scripts for e.g. interfacing data to other systems, the project would have both the source and target database use the same database so only the source side needed to be retrieved for later on generating the desired export scripts. |
|
Migrator - Program Generation
|
|
| Below screenshot shows an excerpt of a generated export script of a different table. From this one can see that a certain naming convention was applied, such as "mig" as prefix, followed by program number and "e" for "export. This short name convention may be required in case of the actual server, responsible for execution of the programs, only allows short names. Another observation would be, that if the below example table were part of our previous exmample project, the mapping for the first field would need to be revised as the "#" sign would not be available or allowed in the target database. Instead of manually changing the code in the generated file, the mapping were to be changed once and any subsequent program generation would contain the correct name. |
|
 |
|
