MAG format schema
This page documents each of the columns and tables in the MAG format data dump, as well as how they all connect to one another.
The MAG format snapshot is for people migrating to OpenAlex from MAG (Microsoft Academic Graph) — you'll want to check out the MAG migration guide for more info.
You may also want to pay particular attention to tables and columns marked ❄️FROZEN (no longer updated) and 🔥NEW (data was not included in original MAG).

Affiliations

Base table for affiliations/institutions (mag/Affiliations.txt)
Field Name
Data Type
Description
AffiliationId
bigint
PRIMARY KEY
Rank
integer
❄️FROZEN
NormalizedName
varchar
UPDATED; slightly different normalization algorithm
DisplayName
varchar
GridId
varchar
❄️FROZEN; RorId is the new standard identifier for organizations
RorId
varchar
🔥NEW; ROR for this organization, see https://ror.org, https://ror.org/:RorId
OfficialPage
varchar
WikiPage
varchar
PaperCount
bigint
PaperFamilyCount
bigint
❄️FROZEN; same value as PaperCount
CitationCount
bigint
Iso3166Code
varchar
Two-letter country codes, see https://en.wikipedia.org/wiki/ISO_3166-2
Latitude
real
Longitude
real
CreatedDate
varchar
UpdatedDate
timestamp
🔥NEW; set values updated from new ror data

AuthorExtendedAttributes

Additional author name representations (mag/AuthorExtendedAttributes.txt)
Field Name
Data Type
Description
AuthorId
bigint
FOREIGN KEY REFERENCES Authors.AuthorId
AttributeType
integer
Possible values: 1=Alternative name
AttributeValue
varchar

Authors

Base table for authors (mag/Authors.txt)
Field Name
Data Type
Description
AuthorId
bigint
PRIMARY KEY
Rank
integer
❄️FROZEN
NormalizedName
varchar
UPDATED; slightly different normalization algorithm
DisplayName
varchar
Orcid
varchar
🔥NEW; ORCID identifier for this author (see https://orcid.org).
LastKnownAffiliationId
integer
FOREIGN KEY REFERENCES Affiliations.AffiliationId
PaperCount
bigint
PaperFamilyCount
bigint
❄️FROZEN; same value as PaperCount
CitationCount
bigint
CreatedDate
varchar
UpdatedDate
timestamp
🔥NEW; set when changes are made going forward

ConferenceInstances

❄️FROZEN; Base table for Conference Instances (mag/ConferenceInstances.txt)
Field Name
Data Type
Description
ConferenceInstanceId
bigint
PRIMARY KEY
NormalizedName
varchar
UPDATED; slightly different normalization algorithm
DisplayName
varchar
ConferenceSeriesId
bigint
FOREIGN KEY REFERENCES ConferenceSeries.ConferenceSeriesId
Location
varchar
OfficialUrl
varchar
StartDate
varchar
EndDate
varchar
AbstractRegistrationDate
varchar
SubmissionDeadlineDate
varchar
NotificationDueDate
varchar
FinalVersionDueDate
varchar
PaperCount
bigint
PaperFamilyCount
bigint
❄️FROZEN; same value as PaperCount
CitationCount
bigint
Latitude
real
Longitude
real
CreatedDate
varchar

ConferenceSeries

❄️FROZEN; Base table for Conference Series (mag/ConferenceSeries.txt)
Field Name
Data Type
Description
ConferenceSeriesId
bigint
PRIMARY KEY
Rank
integer
❄️FROZEN
NormalizedName
varchar
UPDATED; slightly different normalization algorithm
DisplayName
varchar
PaperCount
bigint
PaperFamilyCount
bigint
❄️FROZEN; same value as PaperCount
CitationCount
bigint
CreatedDate
varchar

EntityRelatedEntities

❄️FROZEN; Relationship between papers, authors, fields of study. (advanced/EntityRelatedEntities.txt)
Field Name
Data Type
Description
EntityId
bigint
EntityType
varchar
Possible values: af (Affiliation), j (Journal), c (Conference)
RelatedEntityId
bigint
RelatedEntityType
varchar
Possible values: af (Affiliation), j (Journal), c (Conference)
RelatedType
integer
Possible values: 0 (same paper), 1 (common coauthors), 2 (co-cited), 3 (common field of study), 4 (same venue), 5 (A cites B), 6 (B cites A)
Score
real
Confidence range between 0 and 1. Larger number representing higher confidence.

FieldOfStudyChildren

Relationship between Fields of Study (advanced/FieldOfStudyChildren.txt)
Field Name
Data Type
Description
FieldOfStudyId
bigint
FOREIGN KEY REFERENCES FieldsOfStudy.FieldOfStudyId
ChildFieldOfStudyId
bigint
FOREIGN KEY REFERENCES FieldsOfStudy.FieldOfStudyId

FieldOfStudyExtendedAttributes

Other identifiers for Fields of Study (advanced/FieldOfStudyExtendedAttributes.txt)
Field Name
Data Type
Description
FieldOfStudyId
bigint
FOREIGN KEY REFERENCES FieldsOfStudy.FieldOfStudyId
AttributeType
bigint
Possible values: 1 (AUI from UMLS), 2 (source url), 3 (CUI from UMLS)
AttributeValue
varchar

FieldsOfStudy

Base table for Fields of Study (advanced/FieldsOfStudy.txt)
Field Name
Data Type
Description
FieldOfStudyId
bigint
PRIMARY KEY
Rank
varchar
❄️FROZEN
NormalizedName
varchar
UPDATED; slightly different normalization algorithm
DisplayName
varchar
MainType
varchar
Level
integer
Possible values: 0-5
PaperCount
bigint
PaperFamilyCount
bigint
❄️FROZEN; same value as PaperCount
CitationCount
bigint
CreatedDate
varchar

Journals

Base table for Journals (mag/Journals.txt)
Field Name
Data Type
Description
JournalId
bigint
PRIMARY KEY
Rank
integer
❄️FROZEN
NormalizedName
varchar
UPDATED; slightly different normalization algorithm
DisplayName
varchar
Issn
varchar
UPDATED; the ISSN-L for the journal
Issns
varchar
🔥NEW; JSON list of all ISSNs for this journal (example: '["1469-5073","0016-6723"]' )
IsOa
boolean
🔥NEW; TRUE when the journal is 100% OA
IsInDoaj
boolean
🔥NEW; TRUE when the journal is in DOAJ
Publisher
varchar
Webpage
varchar
PaperCount
bigint
PaperFamilyCount
bigint
❄️FROZEN; same value as PaperCount
CitationCount
bigint
CreatedDate
varchar
UpdatedDate
timestamp
🔥NEW; set when changes are made going forward

PaperAbstractsInvertedIndex

Inverted index of abstracts (nlp/PaperAbstractsInvertedIndex.txt{*} split across multiple files)
Field Name
Data Type
Description
PaperId
bigint
FOREIGN KEY REFERENCES Papers.PaperId
IndexedAbstract
varchar

PaperAuthorAffiliations

Links between papers, authors, and affiliations/institutions. NOTE: It is possible to have multiple rows with same (PaperId, AuthorId, AffiliationId) when an author is associated with multiple affiliations. (mag/PaperAuthorAffiliations.txt)
Field Name
Data Type
Description
PaperId
bigint
FOREIGN KEY REFERENCES Papers.PaperId
AuthorId
bigint
FOREIGN KEY REFERENCES Authors.AuthorId
AffiliationId
bigint
FOREIGN KEY REFERENCES Affiliations.AffiliationId
AuthorSequenceNumber
integer
1-based author sequence number. 1: the 1st author listed on paper, 2: the 2nd author listed on paper, etc.
OriginalAuthor
varchar
OriginalAffiliation
varchar

PaperCitationContexts

❄️FROZEN; citation contexts (nlp/PaperCitationContexts.txt)
Field Name
Data Type
Description
PaperId
bigint
FOREIGN KEY REFERENCES Papers.PaperId
PaperReferenceId
bigint
FOREIGN KEY REFERENCES Papers.PaperId
CitationContext
varchar

PaperExtendedAttributes

Extra paper identifiers (mag/PaperExtendedAttributes.txt)
Field Name
Data Type
Description
PaperId
bigint
FOREIGN KEY REFERENCES Papers.PaperId
AttributeType
integer
Possible values: 1=PatentId, 2=PubMedId, 3=PmcId, 4=Alternative Title
AttributeValue
varchar

PaperFieldsOfStudy

Linking table from papers to fields, with score (advanced/PaperFieldsOfStudy.txt)
Field Name
Data Type
Description
PaperId
bigint
FOREIGN KEY REFERENCES Papers.PaperId
FieldOfStudyId
bigint
FOREIGN KEY REFERENCES FieldsOfStudy.FieldOfStudyId
Score
real
Confidence range between 0 and 1. Bigger number representing higher confidence.
AlgorithmVersion
integer
🔥NEW; version of algorithm to assign fields. Possible values: 1=old MAG (❄️FROZEN), 2=OpenAlex

PaperMeSH

MeSH headings assigned to the paper by PubMed (advanced/PaperMeSH.txt)
Field Name
Data Type
Description
PaperId
bigint
FOREIGN KEY REFERENCES Papers.PaperId
DescriptorUI
varchar
more
DescriptorName
varchar
more
QualifierUI
varchar
more
QualifierName
varchar
more
IsMajorTopic
boolean
more

PaperRecommendations

Paper recommendations with score (advanced/PaperRecommendations.txt)
Field Name
Data Type
Description
PaperId
bigint
FOREIGN KEY REFERENCES Papers.PaperId
RecommendedPaperId
bigint
FOREIGN KEY REFERENCES Papers.PaperId
Score
real
Confidence range between 0 and 1. Bigger number representing higher confidence.

PaperReferences

Paper references and, in reverse, citations (mag/PaperReferences.txt)
Field Name
Data Type
Description
PaperId
bigint
FOREIGN KEY REFERENCES Papers.PaperId
PaperReferenceId
bigint
FOREIGN KEY REFERENCES Papers.PaperId

PaperResources

❄️FROZEN. Data and code urls associated with papers (mag/PaperResources.txt)
Field Name
Data Type
Description
PaperId
bigint
FOREIGN KEY REFERENCES Papers.PaperId
ResourceType
integer
Bit flags: 1=Project, 2=Data, 4=Code
ResourceUrl
varchar
Url of resource
SourceUrl
varchar
List of urls associated with the project, used to derive resource_url
RelationshipType
integer
Bit flags: 1=Own, 2=Cite

PaperUrls

Urls for the paper (mag/PaperUrls.txt)
Field Name
Data Type
Description
PaperId
bigint
FOREIGN KEY REFERENCES Papers.PaperId
SourceType
integer
Possible values: 1=Html, 2=Text, 3=Pdf, 4=Doc, 5=Ppt, 6=Xls, 8=Rtf, 12=Xml, 13=Rss, 20=Swf, 27=Ics, 31=Pub, 33=Ods, 34=Odp, 35=Odt, 36=Zip, 40=Mp3, 0/999/NULL=unknown
SourceUrl
varchar
LanguageCode
varchar
UrlForLandingPage
varchar
🔥NEW; URL for the landing page, when article is free to read
UrlForPdf
varchar
🔥NEW; URL for the PDF, when article is free to read
HostType
varchar
🔥NEW; host type of the free-to-read URL, Possible values: publisher, repository
Version
varchar
🔥NEW; version of the free-to-read URL Possible values: submittedVersion, acceptedVersion, publishedVersion (more)
License
varchar
🔥NEW; license of the free-to-read URL (example: cc0, cc-by, publisher-specific)
RepositoryInstitution
varchar
🔥NEW; name of repository host of URL
OaiPmhId
varchar
🔥NEW; OAH-PMH id of the repository record

Papers

Main data for papers (mag/Papers.txt)
Field Name
Data Type
Description
PaperId
bigint
PRIMARY KEY
Rank
integer
❄️FROZEN
Doi
varchar
Doi values are upper-cased per DOI standard
DocType
varchar
Possible values: Book, BookChapter, Conference, Dataset, Journal, Patent, Repository, Thesis, NULL : unknown. Patent is REMOVED; no patents are included.
Genre
varchar
🔥NEW; Crossref ontology for work type such as journal-article, posted-content, dataset, or book-chapter
IsParatext
boolean
PaperTitle
varchar
UPDATED; slightly different normalization algorithm
OriginalTitle
varchar
BookTitle
varchar
Year
integer
Date
varchar
OnlineDate
varchar
Publisher
varchar
JournalId
bigint
FOREIGN KEY references Journals.JournalId
ConferenceSeriesId
bigint
❄️FROZEN. FOREIGN KEY references ConferenceSeries.ConferenceSeriesId;
ConferenceInstanceId
bigint
❄️FROZEN. FOREIGN KEY references ConferenceInstance.ConferenceInstanceId;
Volume
varchar
Issue
varchar
FirstPage
varchar
LastPage
varchar

RelatedFieldOfStudy

Relationships between fields of study (advanced/RelatedFieldOfStudy.txt)
Field Name
Data Type
Description
FieldOfStudyId1
bigint
FOREIGN KEY REFERENCES FieldsOfStudy.FieldOfStudyId
Type1
varchar
Possible values: general, disease, disease_cause, medical_treatment, symptom
FieldOfStudyId2
bigint
FOREIGN KEY REFERENCES FieldsOfStudy.FieldOfStudyId
Type2
varchar
Possible values: general, disease, disease_cause, medical_treatment, symptom
Rank
real
❄️FROZEN.