2021-07-15 Metadata Management Meeting notes

BugFest is live. Please do jump-in. Ann-Marie asked for more testers around data import template issues; BugFest channel has request.

• PPT in meeting recording
• Some instances have thousands of items; when exporting in custom profile that includes bib, holdings and items, exceeds 99,999 character limit - in which case export fails
  • Workaround: identify record, exclude record and then export that bib separately
• Options
  1. Ignore MARC record limit
     
     1. Pro: all data exported
     2. Con - makes generating MARC record unusable / difficult to use
  2. Truncate data; with or without logging the error
     1. Pro: error in processing of one record does not affect whole export; easiest to implement
     2. Con: no control over which items are exported
     3. Implementation options: export only MARC bib (no holdings & items); log holdings and items removed to allow for separate export with different profile
  3. Export invalid data to a separate file
     1. Pro: clear separation of the records; all data are exported
     2. Con: difficult to implement; requires architectural changes to current implementation to support two files associated with one export
• Use Case for exporting item data with the bib & holdings in .mrc: export for discovery layer;
• Handling in existing systems:
  • Chicago: have split serial records into many to limit the number of items/bib
  • R Scheier: have bib records that exceed the limit themselves (content notes)
  • 5C: items are not extracted... happens via look-up
• Can system ignore AND log those records them? Yes – but then cannot export those... unless they have a separate mapping profile with simplified exported record
• Log file needed regardless!!!
• Approach: if record exceeds limit, will not be included in export but a log will be created indicated excluded records
  • Option 3 is long-term desired solution (as per a few voices)

• Main scope: review UXPROD-2712 and prioritize remains work
• Wiki pages under MM-SIG's working groups
• 30 use cases identified, reviewed and prioritized
• Remaining to do:
  • call number search
  • exact phrase match: contributors, titles, other full-text fields that are analyzed and pre-processed
  • count of matching records; count instances when counting holdings or items may be more appropriate (see later "Inventory Results List Updates" agenda item)
• Supported Search Types: https://github.com/folio-org/mod-search#supported-search-types
• Question(s):
  • Will these enhancements be available in PO search as well? No, only inventory but worth considering for the instance look-up, which is a separate piece of software / plug-in (UIPFI project). Issues at-hand: UI and data source (Inventory storage)
  • Searching across all properties? Discussed during last meeting to allow for searching all already-indexed properties (this does not include all properties); there are some fields deemed unnecessary for search... and if we build search for all edge cases, may lose sight of 99% of searches. All are welcome to join subgroup meetings to engage in discussion; reach out to Magda if you'd like to attend.

Slidedeck

x

VZG

Stanford

Missouri States University

x

Missouri State University

Index Data