MARC records from vendors

Revision as of 18:39, 9 November 2022 by ErinBlake (talk | contribs) (→‎Gale databases: Finished ECCO notes for BSLW sample)

MARC records supplied by vendors require editing before they can be batch-loaded into the catalog. This page describes edits that need to be made to all vendor-supplied records.

Procedures for all formats

  1. Break file from .mrc to .mrk
  2. Make back-up copy of unedited .mrk file
  3. Open working .mrk file in MarcEditor
    1. Check that resources list of changes, below, against the content to make sure they're still accurate and relevant
    2. Make any necessary adjustments to the tasks in Tools > Manage tasks > Task Actions > Edit task > [Dataset name]
    3. Run Tools > Assigned tasks > [Dataset name]
    4. Run any other relevant MarcEdit tasks (e.g. replacing degree sign in ESTC formats)
    5. Save as MARC21 XML
  4. Open in Notepad++
    1. Run Macroexpress macro to replace LDR with 000 and control field spaces with backslashes
    2. If replacing existing records, run Macroexpress macros to delete the 005
  5. Batch upload in TIND
    1. For large record sets, first use MarcEdit to split file into sets of about 1,000 records each
    2. Schedule uploads half an hour apart

Additional procedures for printed resources

Fields to add

  • 500__ $a This record was provided by a vendor. It may contain incorrect or incomplete information. $5 DFo
  • 980__ $a BIB
  • 983__ $a Not-vault

Additional procedures for online facsimiles

Delete:

  • 506 Restrictions on access (Folger reserves this for physical restrictions; license restrictions are implicit in 852$h)
  • 538 ("Mode of access: Internet" and "Mode of access: WWW" are no longer necessary)

Replace, if necessary:

  • LDR/06: do not use "m" for digital or digitized documents (electronic aspects are coded in 006 and 007 instead)
  • LDR/06: for digitized manuscripts, use codes t, d, and f (Folger practice is contrary to OCLC, which considers digitized manuscripts to be "published" resources)
  • 008/23: use "o" ("online") not "s" ("electronic")
  • 006 for digital and digitized documents: m||||o|||d||||||||
  • 007 for digital scans of documents: cr\un|---uuuua (in theory, "color" and "antecedent/source" could be coded something other than "unknown" but in practice, we want the records to remain valid if a vendor replaces page images scanned from microfilm with page images scanned from the original)
  • 007 for native ebooks: cr\un|---uuuun
  • Link text in 852 other than $u if they don't make sense for the Folger catalog

Move:

  • Vendor 001 and 003 move to 035, in this format and order (unless a different match point is required): (003)001

Add:

 245%% $$h[electronic resource]
 336__ $$atext$$btxt$$2rdacontent
 337__ $$acomputer$$bc$$2rdamedia
 338__ $$aonline resource$$bcr$$2rdacarrier
 500__ $$aThis record was provided by a vendor. It may contain incorrect or incomplete information.$$5 DFo
 5831_ $$abatch edited$$cyyyy-mm-dd$$kEB$$xMarcEdit Task$$2local$$5DFo
 5831_ $$abatch loaded$$cyyyy-mm-dd$$kEB$$xfrom edited vendor records$$2local$$5DFo
 852__ $$aUS-DFo$$bEleRes$$hAvailable offsite via https://request.folger.edu
 980__ $$aBIB
 983__ $$aOnline

Don't worry about:

  • 040$d (because editing by Folger staff is automatically recorded in TIND record history and manually recorded in MARC 583)

Resources with vendor-supplied record sets

ACLS Humanities E-Book collection

  • 5,472 records (plus 2 duplicates)
  • Remove duplicates from original .mrc file based on 035
  • Run MARCValidator on deduped source file to make sure nothing more needs to be added to the MarcEdit tasks
    • Don't worry about the repeated 010 fields
  • Delete 776$q
  • Replace =830 \\$aACLS with =830 \0$aACLS
  • Change =830 \0$aATLA Special Series. 0$aACLS Humanities E-Book. to have a separate 830 for $aACLS Humanities E-Book
  • Change =830 \0$aATLA Special Series. 0$aAmerican philosophy series ;$vno. 7. to have a separate 830 for American philosophy series
  • Delete MARC 506
  • Delete MARC 538
  • Delete MARC 773
  • Replace LDR/07 d or a with m
  • Change 008/23 (form of item) from "s - electronic" to the more specific "o - online"
  • Replace 006 with m||||o|||d||||||||
  • Replace 007 with cr\un|---uuuun (most are okay as-is, but about two dozen have too many characters in the string)
  • Delete 856$z
  • Add RDA 33X fields:
    • 336__ $a text $b txt $2 rdacontent
    • 337__ $a computer $b c $2 rdamedia
    • 338__ $a online resource $b cr $2 rdacarrier
  • Add standard HBCN, 583, 980, 983
  • Add 852__ $$aUS-DFo$$bEleRes$$hAvailable onsite only but use the list of Open Access titles to batch change the $h to Open Access after records are loaded.

Adam Matthew databases

Available from https://www.amdigital.co.uk/support/marc-records, login not required.
Products with updated records (i.e. updated since the initial release) will include an update date in the title. The zipped folder will contain the updated records.
For updates made after May 2019, the zipped folder will contain the updated records as well as an Update Notes text file detailing the changes.
NOTE: As discussed with Researcher Services, we currently do not upload records for individual titles in Adam Matthew resources, just the record for the resource itself (the individual records are too non-standard to be helpful).

  • Early Modern England: 772 records (none uploaded)
    • LDR always coded a, p, or r, never "t", so manuscripts can only be determined from 245$k such as "Manuscripts, essays, memoirs and music" or "Commonplace books; Essays" (which the database uses to generate the "Document type" field)
    • All have encoded date nuuu and imprint "Marlborough, Wiltshire : Adam Matthew Digital" (except for seven that also have parts of the original imprint in the 264).
    • 6XX fields are all non-standard (except for 653, which is just keywords)
  • Literary Manuscripts Leeds: 190 records (none uploaded)
    • LDR is mix of "a" and "t" (though originals are all manuscripts)
    • Fixed field date always coded for the database itself
    • Imprint is always for the database itself
    • 300 is always for the original manuscript
    • Most 650s should be 655s (and don't have a $v)
  • Literary Print Culture: 1,908 records (none uploaded)
    • LDR is always "a"
    • Fixed field date always coded for the database itself
    • Imprint is for the database itself
    • 300 is always "1 online resource"
    • Nothing to indicate they're manuscripts (245$k is a category, e.g. =245 00$aBond following illegal publication of almanacs :$kLegal record; Financial record$g1723.
    • Keywords in 653 provide the only 6XX access
  • Perdita Manuscripts, 1500-1700: 216 records (none uploaded)
    • LDR varies: a, p, t (but all are manuscripts)
    • 008 and imprint are for the online resource
    • 300 is "1 online resource"
    • 6XX are non-standard (plus 653 with keywords)
  • Shakespeare in performance: 1,138 records (none uploaded)
    • LDR always "a" (material includes manuscripts, graphic materials, published texts)
    • 008, 264, and 300 are all for the electronic resource
    • 245$k provides the general category (e.g. "Photograph", "Costume design", "Autograph; Manuscript; Letters")
    • 534 correctly has "$p Reproduction of:" but dates in $c are inconsistent, and ISBD punctuation is lacking
    • Keywords in 653 provide the only 1XX, 6XX, or 7XX access
  • Shakespeare's Globe Archive: 1,759 records for documents, 26 records for non-musical recordings (none uploaded)
    • LDR/06 is "i" for sound files (oral history interviews) and "a" for everything else (incl. props, graphic materials, manuscripts)
    • 008, 264, and 300 are for the online resource
    • 100 has only $a and $e
    • 534 correctly has "$p Reproduction of:" but dates in $c are inconsistent, and ISBD punctuation is lacking
    • The only 7XX is a constant "710 2\$aAdam Matthew Digital (Firm),$edigitiser."
  • Virginia Company Archives: 2,899 records (none uploaded)
    • LDR/06 is always "t"
    • 008, 264, and 300 are all for the online resource
    • In all but about 20 records, the 245$a is the call number, 245$k is "Manuscripts" (even when item is a printed picture).
    • 520 is used for the item title
    • 6XX is non-standard

Gale databases

  • (by request, except for British Literary Manuscripts)
  • British Literary Manuscripts
    • from https://support.gale.com/marc/, login not required)
    • 17 records for source microfilm collection titles (leads to description and targeted search)
    • 7 records for targeted searches for each of seven parts of the Medieval literary and historical manuscripts in the Cotton Collection, British Library, London (links break after the tilde when followed in TIND ILS, but can be copy-and-pasted from the MARC)
    • In addition to Online facsimiles changes described above:
      • Delete 538
      • Replace =[LOCATIONID] with =wash46354
      • Replace existing 260 (which is for the microfilm) with 260 for Gale Cengage Learning, coded as current/latest publisher (that is, 260 3_ $a [Farmington Hills, Michigan] : $b Gale Cengage Learning, $c [2009?]). Note: Gale Cengage Learning now rebranded as Gale a Cengage Company, but BLM hasn't been updated (yet).
      • 008 change Date 1 to 2009, change 008/15-17 to miu, change 008/23 from "s" to "o"
      • 006 and 007 are okay as-is
      • Replace any existing 300 with 300 $a 1 online resource; add same to records that lack a 300
  • Burney Collection Newspapers
    • 1,057 records
    • Must transform from MARC8 to UTF8
    • Already has the wash46354 suffix in URL
    • Remember that many 008s are coded for serials, not books
    • Delete the one instance of ";$c?°." then run MarcEdit Degree sign format fix
    • Remove duplicate 035s from source file (e.g. 35 records with =035\\$ageneral)
    • Change 006 to m||||o|||d||||||||
    • Change 007 to cr\un|---uuuua
    • Change 008/23 from "s" to "o"
    • Add 33X
    • Delete any existing 533$n
    • Update 520s for STC, Wing, and ESTC?
    • Move 590$a to 533$n
    • Delete 648
    • Change 985 to 509
  • Eighteenth century collections online
    • 184,371 records in three separate files
    • Must transform from MARC8 to UTF8
    • Already has the wash46354 suffix in URL
    • Remember that many 008s are coded for serials, not books
    • Need to return to this after BSLW processing; so far only did six test records rather than check the whole file for errors and anomalies
      • Would normally run the MarcEdit Degree sign format fix, but opting not to for the small batch of test records in order to verify that BSLW's script is still working.
    • Delete existing 852s and replace with FSL custom
    • Change 006 to m||||o|||d||||||||
    • Change 007 to cr\un|---uuuua
    • Make 008/23 be "o" (unless LDR coded for Maps or Visual Materials, in which case 008/29 = "o")
    • Delete any existing 533$n
    • Deliberately did NOT add 33X fields or update 510s
    • Standard TIND transformations
  • Nichols Newspapers Collection
    • 660 records
  • British Theatre, Music, and Literature: High and Popular Culture (from Nineteenth Century Collections Online)
    • 608 records

Loeb Classical Library

Oxford Scholarly Editions Online (OSEO)

  • 272 records
  • Remove individual records for each volume of Martin Wiggins and Catherine Richardson, British Drama 1533–1642: A Catalogue in favor of one record for the entire set (rec id 544900): subject headings are identical, and tables of contents are very incomplete (e.g. v. 1 covers 1533-1566 with entries for 440 plays, but table of contents only has entries for 37 plays, and goes to the start of 1537). Note: punctuation in 245$a varies, so search by field "F#:245$c Martin Wiggins and Catherine Richardson (eds)" to find all nine in MarcEdit.
  • Note: MARC 008 for both volumes of The correspondence of Sir Philip Sidney is incorrect. Vendor record has 130909s2012\\\\enka\\\fo|\\\o0|0\0\eng|d but should have 130909s2012\\\\enka\\\fob\\\\001\0\eng|d.

ProQuest databases

  • Log in to https://admin.proquest.com/requestmarcrecords to get records via FTP
  • Early English books online (EEBO): 133,109 records, with updates annually in November
    • Problems to ignore (for now)
      • Thousands of records have entire 510 in $c (planning to split out $c only the 8 that are for STCs: all begin "=510 4\$aSTC (2nd ed.) " [fixed]
      • Too many variations in 510$a to do a comprehensive fix without OpenRefine. Ran the two scripts developed for Burney and left it at that for now.
      • Dozens of records have Date2 in the range between 1899 and 2005 as the latest possible date for Date1 of "some time after x"
    • Convert to UTF8
    • Run MarcEdit macro to change 4° to 4to, etc. [done] and then replace degree signs representing superscript "o" with "o" on-the-line (per DCRM rules for superscript abbreviations) [done]
    • Change 008/23 from blank to "o"
    • Change LDR/18 from I to i
    • Add MARC 852 \\$aUS-DFo$bEleRes$hAvailable offsite via https://request.folger.edu
    • Add MARC 830 \0$aEarly English books online. if not already present
    • Append /?accountid=10923 to every 856
    • Delete 648
    • Add 336 when not already present
    • Delete any existing 337
    • Add 337 $a computer
    • Delete any existing 338
    • Add 338 for online resource
    • Use 006 for digital and digitized documents: m||||o|||d||||||||
    • Use 007 for digital scans of documents: cr\un|---uuuua
    • Replace all instances of DFo and DFO in MARC 500 with "Folger Shakespeare Library"
    • Replace instances of "$Madan, F. Oxford books" in 510 with "$aMadan, F. Oxford books"
    • Replace instances of "$Union Theological Seminary" in 510 with "$aUnion Theological Seminary"
    • Use 001 to build new 035 in the form (DFo_eebo_001) to match existing imported eebo records, then delete 001 and 003
    • Delete 6XX $2fast headings
    • Delete 655 $2gsafd headings
    • Delete 655 $2lcgft headings
    • Change $aGreat Britian to $aGreat Britain (there are other instances of "Britian" but they're in notes so not worth bothering with)
    • Fix problematic 752s (use exact word match to avoid overwriting partial strings)
      • $aEngland and Wales$dLondon > $aGreat Britain$bEngland$dLondon
      • $aEngland > $aGreat Britain$bEngland
      • $aEnglan > $aGreat Britain$bEngland
      • $aGreat Britain$bIreland > $aIreland
      • $aIreland$dLondonderry > $aGreat Britain$dLondonderry
      • $aScotland > $aGreat Britain$bScotland
      • $aUnited States$bNew York$dNew York > $aUnited States$bNew York (State)$dNew York
    • Delete subdivisions v, x, y, and z in 655, then add back the missing dot using regex find "(=655 \\7\$a.*)([^.])(\$2rbgenr)" replace with "$1s.$3" (Note: assumes all rbgenr $a terms end with "s", only needs to be done for rbgenr b/c others have closing parenthesis as terminal punctuation)
    • Change =656 to =655
    • Change 956 to 500
    • Delete MARC 871 field (there's no such thing: it's a data error)
    • Delete MARC 952 (contains an "eebo" code, conflicts with Folger use of 952)
    • Delete MARC 520 if is "\\$aeebo-0018" or "\\$aeebo-0216"
    • Delete MARC 994
    • Delete MARC 999
    • Change date in 008 from 15001599 to 15uu\\\\ (or else material with 260$c [15--?] will be pulled into the Incunabula collection)
    • Country codes to fix:
      • eng to enk
      • ir to ie
      • ng to ne
    • Language codes to fix
      • change N/A to \\\
      • change gae to gla
      • change iri to gle
    • Known problems with individual records to fix by hand:
      • Find and fix 20 records where first character of Date1 isn't "1" =008.{9}[^1] [fixed]
      • Manuscript fragments coded as books:
        • useful regex: fragment.*manuscript [done]
        • phrase search: "Catalogue of Manuscripts containing Anglo-Saxon" (no need to search "Additional mss" or "Harleian mss" b/c those are already correct) [done]
      • The tell-truth remembrancer dates should be d17021703 and lang should be eng [fixed]
      • EEBO rec id 12928949 has 510$c 0574 instead of O574 [fixed]
      • EEBO rec id 250600827 has 510$a that also contains huge part of the table of contents [fixed]
      • Two records have a 510 beginning " Wing" [fixed]
      • remove 16 non-authorized descriptive name main entries that start with "A" (e.g., "A Gentleman of Good Quallity") TIND search= "1000%a:/^A /" Notepad++ literal search= "=100 0\$aA " (don't worry about the hundreds of others: they won't jump to the top of the name facet as a group) [fixed]

List of resources without vendor records for individual titles

  • Gale databases
    • State papers online : the government of Britain, 1509-1714.
  • ProQuest databases
    • Cecil papers