Rerun Solr export and Fedora query for Disk Image objects to create object-level metadata files

[kmichaelis]

• Rerun Solr export and query Fedora as needed to make sure that all Keep metadata fields are included in the exported file (some fields not indexed in Solr).

• Please place exported CSV in the project folder.

• Split the exported file into object-level metadata files in whatever format is easiest (e.g. txt, csv)

• Add each object-level metadata file to the folder containing the object it describes. Folders should be named with the object's pid.

[alexBLR]

@kmichaelis The largest collections by size:
emory:94k9k (Salman Rushdie papers) — 12 files, 587 GB
emory:tgn0r — 64 files, 452 GB
emory:94kf4 — 12 files, 347 GB
emory:gk0st — 3 files, 266 GB
emory:f3qwd — 8 files, 250 GB
File counts range from 1 to 316 per collection, and sizes from a few KB up to 206 GB for a single object.

[kmichaelis]

KM and KF confirm whether the S3 issue is impacting this

[rotated8]

@kmichaelis @katfish2 Please review this export

[kmichaelis]

@rotated8 I'm unable to access the export--see attached error message

[rotated8]

@kmichaelis Try this link

[kmichaelis]

Thanks @rotated8 , this one works. @katfish2

[katfish2]

That looks like the export of all Solr fields. I don't see any problems after a quick review.

We do still need the additional info from Fedora before the info is split into object-level metadata files and ingested. Is that a separate export? There are a few essential fields that aren't in the Solr index—can someone track down the list in the R&R project documentation? I never had access to those tickets.

[kmichaelis]

These are the fields

[rotated8]

Here's the report with the pids missing from the Fedora data: Pid Matching

[AGCooper]

The promised sample file (sans crosswalk) can be found here under Collections and Objects.