The DocStore is where bibliographic data such as MARC records, holdings and items are stored in OLE.
Institutions typically have millions of MARC data to be loaded for the very first time when they implement OLE. It is impossible to use conventional methods like Batch process to load the data considering both time and memory constraints.
A simpler method is to directly load the MARC data as MARCXML into the database and then do a re-indexing. The whole process is extensively documented here.
However, Institutions typically receive MARC data in .mrc or .out (Millennium) formats from Library Management Systems (LMS). This can be converted to MARCXML format with help of third-party tools such as MARCEdit for Windows, YAZ for Linux or other tools recommended by Library of Congress. The Institution should ensure the integrity of data so converted as certain tools are known to mess data owing to encoding problems (A version of YAZ always ignored or replaced diacritic characters).
Holdings and Items data
The 9XX fields are earmarked for institution specific data in MARC and this is widely leveraged by LMS to carry Item and Holding information. Export of MARC data from legacy LMS may not be complete. In Millennium, for example, the MARC records extracted lacked Item Bar code and Status information and necessitated a special plugin from Innovative Interfaces, Inc (III) to extract the same.