D-Net
/
dnet-hadoop
14
5
Fork 11
Code Issues 14 Pull Requests 13 Projects 2 Releases 2 Wiki Activity

[doiboost] 

- update MAG schema, new filed added on version dec-2021
This commit is contained in:
Sandro La Bruzzo 2022-01-11 11:05:44 +01:00
parent 904e1c2667
commit 0163dadb7f
1 changed files with 1 additions and 1 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

2
dhp-workflows/dhp-doiboost/src/main/scala/eu/dnetlib/doiboost/mag/SparkImportMagIntoDataset.scala
View File

@ -41,7 +41,7 @@ object SparkImportMagIntoDataset {
"PaperReferences" -> Tuple2("mag/PaperReferences.txt", Seq("PaperId:long", "PaperReferenceId:long")), "PaperReferences" -> Tuple2("mag/PaperReferences.txt", Seq("PaperId:long", "PaperReferenceId:long")),
"PaperResources" -> Tuple2("mag/PaperResources.txt", Seq("PaperId:long", "ResourceType:int", "ResourceUrl:string", "SourceUrl:string", "RelationshipType:int")), "PaperResources" -> Tuple2("mag/PaperResources.txt", Seq("PaperId:long", "ResourceType:int", "ResourceUrl:string", "SourceUrl:string", "RelationshipType:int")),
"PaperUrls" -> Tuple2("mag/PaperUrls.txt", Seq("PaperId:long", "SourceType:int?", "SourceUrl:string", "LanguageCode:string")), "PaperUrls" -> Tuple2("mag/PaperUrls.txt", Seq("PaperId:long", "SourceType:int?", "SourceUrl:string", "LanguageCode:string")),
"Papers" -> Tuple2("mag/Papers.txt", Seq("PaperId:long", "Rank:uint", "Doi:string", "DocType:string", "PaperTitle:string", "OriginalTitle:string", "BookTitle:string", "Year:int?", "Date:DateTime?", "OnlineDate:DateTime?", "Publisher:string", "JournalId:long?", "ConferenceSeriesId:long?", "ConferenceInstanceId:long?", "Volume:string", "Issue:string", "FirstPage:string", "LastPage:string", "ReferenceCount:long", "CitationCount:long", "EstimatedCitation:long", "OriginalVenue:string", "FamilyId:long?", "FamilyRank:uint?", "CreatedDate:DateTime")), "Papers" -> Tuple2("mag/Papers.txt", Seq("PaperId:long", "Rank:uint", "Doi:string", "DocType:string", "PaperTitle:string", "OriginalTitle:string", "BookTitle:string", "Year:int?", "Date:DateTime?", "OnlineDate:DateTime?", "Publisher:string", "JournalId:long?", "ConferenceSeriesId:long?", "ConferenceInstanceId:long?", "Volume:string", "Issue:string", "FirstPage:string", "LastPage:string", "ReferenceCount:long", "CitationCount:long", "EstimatedCitation:long", "OriginalVenue:string", "FamilyId:long?", "FamilyRank:uint?", "DocSubTypes:string", "CreatedDate:DateTime")),
"RelatedFieldOfStudy" -> Tuple2("advanced/RelatedFieldOfStudy.txt", Seq("FieldOfStudyId1:long", "Type1:string", "FieldOfStudyId2:long", "Type2:string", "Rank:float")) "RelatedFieldOfStudy" -> Tuple2("advanced/RelatedFieldOfStudy.txt", Seq("FieldOfStudyId1:long", "Type1:string", "FieldOfStudyId2:long", "Type2:string", "Rank:float"))
) )