Importing 7.x/8.0 Libraries

UnForm 10.0 uses a different library structure, and different naming conventions, than previous UnForm versions. A simple command line is used to import an old library into a new, version 10.0 library. By using different naming conventions, it is actually possible for version 8 and 10 libraries to coexist in the same physical location, though it is important to realize that the two libraries do not interact and are not updated in tandem by either version. They are functionally two different libraries, even if they share the same disk path.

Care should be taken when importing, as UnForm 10.0 libraries require more disk space due to new indexing and recovery transaction logs. How much more space required is unpredictable, but it isn't uncommon for disk use to approximately double from previous versions. It may be preferable to create the 10.0 libraries on a different file system, or a different machine altogether, though the import process needs access to both the old and new libraries, so a copy could be made, or network file system access used, to read the old library.

Imports are incremental, in that if a give doctype/docid or doctype/docid/subid already exists in the new library, it will not be overwritten by a subsequent import. This allows importing multiple times, incrementally adding new documents and images. If you want to ensure that an import captures all documents, including those that may have been imported previously and later updated in the older system, you must import into a new, empty library.

uf100c -arclogin "user/password" -arclib "new library" -arcimport "old library" -o "logfile.txt" [-arcdoctype "doctype" | -arcdtm "yyyymmdd[-yyyymmdd]" ]

An administrative login is required to perform the import. The import will run in background, writing logging details to the server file specified in the -o option. Be sure not to specify a client-side output file, or else the client will continue running for as long as the import runs.

Libraries are not locked during the import, so it is possible to use them while the import is running. There are two caveats, however, related to potential duplicate ID values.

•	Any documents that have auto-generated document ID values could cause conflicts, as new documents being added would likely receive an ID that will be duplicated by the import, resulting in skipping the import of the older document.

•	Any sequenced sub ID value is potentially in conflict with a subsequent imported sub ID, resulting in skipping of importing the older image.

If a library contains documents with auto sequencing, either of doc ID's or sub ID's, it is best to not print or upload such documents until the import is complete.

When importing, in addition to the output log file, the system logs entries to the server log at the start and end, and every 15 minutes of the job, to indicate status.

You can specify either a document type or a date/date range by adding -arcdoctype or -arcdtm options to the command line. If a document type is specified, just documents of that type are imported into the new library. Likewise, if you specify a date or date range, just documents in that range of document date are imported. The date can be specified as a year (i.e. 2014), or a month (201401), or a full day (20140115). If you require a more specific range, use a hyphen to specify it (i.e. 20140115-20140215).

By using ranges, you can run multiple import jobs concurrently to speed import processing, within the resource limits of the system.