From 0163dadb7f9abbe1c801b55d30728c4e3768627f Mon Sep 17 00:00:00 2001 From: Sandro La Bruzzo Date: Tue, 11 Jan 2022 11:05:44 +0100 Subject: [PATCH] [doiboost] - update MAG schema, new filed added on version dec-2021 --- .../eu/dnetlib/doiboost/mag/SparkImportMagIntoDataset.scala | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/dhp-workflows/dhp-doiboost/src/main/scala/eu/dnetlib/doiboost/mag/SparkImportMagIntoDataset.scala b/dhp-workflows/dhp-doiboost/src/main/scala/eu/dnetlib/doiboost/mag/SparkImportMagIntoDataset.scala index d25a4893f..039c935f3 100644 --- a/dhp-workflows/dhp-doiboost/src/main/scala/eu/dnetlib/doiboost/mag/SparkImportMagIntoDataset.scala +++ b/dhp-workflows/dhp-doiboost/src/main/scala/eu/dnetlib/doiboost/mag/SparkImportMagIntoDataset.scala @@ -41,7 +41,7 @@ object SparkImportMagIntoDataset { "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")), "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")) )