I have done many migrations where the source data is in a proprietary system or where the team is managing their documents and revisions in an excel spreadsheet. In these cases, I typically receive the Excel workbook as the source metadata and a file system for the attachments. Attachment names often begin with the item number or have a path in the spreadsheet to the directory the file are located in.
For our demonstration I will use a source spreadsheet with a single row of metadata and a file system where the attachment names begin with the item number. From this we can prepare the metadata for the Oracle DataLoad tool and also parse out the file system and build a manifest for the File Load tool (Java Client).
Here are the steps I will take for the migration:
- Import the spreadsheet into an Oracle database. Again, imagine I am migrating 10,000 documents with different release dates and a more complex data set.
- Run SQL against the imported spreadsheet for some clean up, validation and loading into the DataLoad schema
- Validate data in DataLoad
- Perform the Migration
- Import the attachment listing into Oracle
- Run SQL to build a file load manifest and export to a CSV file
- Run File Load to import attachments
In the next post, I will walk through the first four steps to migrate the metadata. I will not be going over the DataLoad tool in depth but will show it and discuss its basic functionality so Agile customers can understand how it works at a high level.
Please do not ask for a copy of the DataLoad tool; you will need to engage OCS or an Agile PLM partner.