Merge #1964 from remote-tracking branch 'ssh/1799-SerialUselessPublDate'

hbz · Feb 6, 2024 · cea402e · cea402e
2 parents a3e819f + 5668b42
commit cea402e
Show file tree

Hide file tree

Showing 4 changed files with 260 additions and 1 deletion.
diff --git a/src/main/resources/alma/fix/titleRelatedFields.fix b/src/main/resources/alma/fix/titleRelatedFields.fix
@@ -342,8 +342,15 @@ do list(path:"533  ", "var": "$i")
 end
 
 # Cleaning up dates of publicaton []
+copy_field("leader","@leaderPos07")
+substring("@leaderPos07","7","1")
 do list(path: "publication[]", "var": "$i")
   replace_all("$i.startDate", ".*?([01]\\d{3}|20\\d{2}).*", "$1")
+  if any_match("@leaderPos07","s")
+    if any_contain("$i.startDate","0000")
+      remove_field("$i.startDate")
+    end
+  end
   replace_all("$i.endDate", ".*-[ ]?([01]\\d{3}|20\\d{2})$", "$1")
   replace_all("$i.location[].*", "^\\[(.*)\\]$", "$1")
   replace_all("$i.location[].*", "\\s?[,:;]$", "")

diff --git a/src/test/resources/alma-fix/990119186660206441.json b/src/test/resources/alma-fix/990119186660206441.json
@@ -0,0 +1,131 @@
+{
+  "@context" : "http://lobid.org/resources/context.jsonld",
+  "almaMmsId" : "990119186660206441",
+  "oclcNumber" : [ "1057478173" ],
+  "zdbId" : "2102077-2",
+  "dnbId" : "024334561",
+  "deprecatedUri" : "http://lobid.org/resources/HT013595767#!",
+  "hbzId" : "HT013595767",
+  "title" : "Know your world extra",
+  "publication" : [ {
+    "type" : [ "PublicationEvent" ],
+    "location" : [ "Middletown, CT" ],
+    "publishedBy" : [ "Xerox Education Publications" ]
+  } ],
+  "describedBy" : {
+    "id" : "http://lobid.org/resources/990119186660206441",
+    "label" : "Webseite der hbz-Ressource 990119186660206441",
+    "type" : [ "BibliographicDescription" ],
+    "inDataset" : {
+      "id" : "http://lobid.org/resources/dataset#!",
+      "label" : "lobid-resources – Der hbz-Verbundkatalog als Linked Open Data"
+    },
+    "resultOf" : {
+      "type" : [ "CreateAction" ],
+      "endTime" : "0000-00-00T00:00:00",
+      "instrument" : {
+        "id" : "https://github.com/hbz/lobid-resources",
+        "type" : [ "SoftwareApplication" ],
+        "label" : "Software lobid-resources"
+      },
+      "object" : {
+        "id" : "https://lobid.org/marcxml/990119186660206441",
+        "dateCreated" : "2021-04-05",
+        "dateModified" : "2023-05-25",
+        "type" : [ "DataFeedItem" ],
+        "label" : "hbz-Ressource 990119186660206441 im Exportformat MARC21 XML",
+        "inDataset" : {
+          "id" : "https://datahub.io/dataset/hbz_unioncatalog",
+          "label" : "hbz_unioncatalog"
+        },
+        "sourceOrganization" : {
+          "id" : "http://lobid.org/organisations/DE-600#!",
+          "label" : "Zeitschriftendatenbank (ZDB)"
+        },
+        "provider" : {
+          "id" : "http://lobid.org/organisations/DE-101#!",
+          "label" : "Deutsche Nationalbibliothek"
+        },
+        "modifiedBy" : [ {
+          "id" : "http://lobid.org/organisations/DE-600#!",
+          "label" : "Zeitschriftendatenbank (ZDB)"
+        } ]
+      }
+    },
+    "license" : [ {
+      "id" : "http://creativecommons.org/publicdomain/zero/1.0",
+      "label" : "Creative Commons-Lizenz CC0 1.0 Universal"
+    } ]
+  },
+  "sameAs" : [ {
+    "id" : "http://worldcat.org/oclc/1057478173",
+    "label" : "OCLC Ressource"
+  }, {
+    "id" : "http://ld.zdb-services.de/resource/2102077-2",
+    "label" : "ZDB-Ressource"
+  }, {
+    "id" : "https://d-nb.info/024334561",
+    "label" : "DNB-Ressource"
+  } ],
+  "fulltextOnline" : [ {
+    "label" : "EZB",
+    "id" : "http://www.bibliothek.uni-regensburg.de/ezeit/?2102077"
+  } ],
+  "related" : [ {
+    "note" : [ "Druck-Ausgabe" ],
+    "label" : "Know your world extra",
+    "issn" : "0163-4844"
+  } ],
+  "inCollection" : [ {
+    "id" : "http://lobid.org/organisations/DE-655#!",
+    "label" : "hbz - Hochschulbibliothekszentrum des Landes Nordrhein-Westfalen, Netzwerkzone",
+    "type" : [ "Collection" ]
+  }, {
+    "id" : "http://lobid.org/resources/HT014846970#!",
+    "label" : "Zeitschriftendatenbank (ZDB)",
+    "type" : [ "Collection" ]
+  }, {
+    "id" : "http://lobid.org/resources/HT016356466#!",
+    "label" : "Elektronische Zeitschriftenbibliothek (EZB)",
+    "type" : [ "Collection" ]
+  } ],
+  "language" : [ {
+    "id" : "http://id.loc.gov/vocabulary/iso639-2/eng",
+    "label" : "Englisch"
+  } ],
+  "extent" : "Online-Ressource",
+  "natureOfContent" : [ {
+    "label" : "Zeitschrift",
+    "id" : "https://d-nb.info/gnd/4067488-5"
+  } ],
+  "subject" : [ {
+    "notation" : "050",
+    "type" : [ "Concept" ],
+    "source" : {
+      "label" : "DDC-Sachgruppen der ZDB"
+    },
+    "label" : "Zeitschriften, fortlaufende Sammelwerke"
+  }, {
+    "type" : [ "Concept" ],
+    "source" : {
+      "label" : "Dewey-Dezimalklassifikation",
+      "id" : "https://d-nb.info/gnd/4149423-4"
+    },
+    "label" : "Allgemeine fortlaufende Sammelwerke",
+    "notation" : "050",
+    "version" : "sdnb"
+  } ],
+  "medium" : [ {
+    "label" : "Datenträger",
+    "id" : "http://rdaregistry.info/termList/RDAMediaType/1003"
+  }, {
+    "label" : "Online-Ressource",
+    "id" : "http://rdaregistry.info/termList/RDACarrierType/1018"
+  } ],
+  "bibliographicLevel" : {
+    "label" : "Serial",
+    "id" : "https://www.loc.gov/marc/bibliographic/bdleader.html#Serial"
+  },
+  "type" : [ "BibliographicResource", "Periodical" ],
+  "id" : "http://lobid.org/resources/990119186660206441#!"
+}
diff --git a/src/test/resources/alma-fix/990119186660206441.xml b/src/test/resources/alma-fix/990119186660206441.xml
@@ -0,0 +1,121 @@
+<?xml version="1.0" encoding="UTF-8"?>
+<record>
+  <leader>00000nas a2200000#c 4500</leader>
+  <controlfield tag="005">20230525135732.0</controlfield>
+  <controlfield tag="007">cr||||||||||||</controlfield>
+  <controlfield tag="008">030218c00009999xxuu||p|o ||| 0||||1eng c</controlfield>
+  <controlfield tag="003">DE-605</controlfield>
+  <controlfield tag="001">990119186660206441</controlfield>
+  <datafield tag="016" ind1="7" ind2=" ">
+    <subfield code="a">2102077-2</subfield>
+    <subfield code="2">DE-600</subfield>
+  </datafield>
+  <datafield tag="016" ind1="7" ind2=" ">
+    <subfield code="a">024334561</subfield>
+    <subfield code="2">DE-101</subfield>
+  </datafield>
+  <datafield tag="035" ind1=" " ind2=" ">
+    <subfield code="a">(DE-599)ZDB2102077-2</subfield>
+  </datafield>
+  <datafield tag="035" ind1=" " ind2=" ">
+    <subfield code="a">(OCoLC)1057478173</subfield>
+  </datafield>
+  <datafield tag="035" ind1=" " ind2=" ">
+    <subfield code="a">(DE-600)2102077-2</subfield>
+  </datafield>
+  <datafield tag="035" ind1=" " ind2=" ">
+    <subfield code="a">(DE-101)024334561</subfield>
+  </datafield>
+  <datafield tag="040" ind1=" " ind2=" ">
+    <subfield code="a">9001</subfield>
+    <subfield code="b">ger</subfield>
+    <subfield code="c">DE-101</subfield>
+    <subfield code="d">9001</subfield>
+  </datafield>
+  <datafield tag="041" ind1=" " ind2=" ">
+    <subfield code="a">eng</subfield>
+  </datafield>
+  <datafield tag="044" ind1=" " ind2=" ">
+    <subfield code="c">XD-US</subfield>
+  </datafield>
+  <datafield tag="084" ind1=" " ind2=" ">
+    <subfield code="a">050</subfield>
+    <subfield code="q">DE-600</subfield>
+    <subfield code="2">sdnb</subfield>
+  </datafield>
+  <datafield tag="245" ind1="0" ind2="0">
+    <subfield code="a">Know your world extra</subfield>
+  </datafield>
+  <datafield tag="264" ind1="3" ind2="1">
+    <subfield code="a">Middletown, CT</subfield>
+    <subfield code="b">Xerox Education Publications</subfield>
+    <subfield code="c">0000-</subfield>
+  </datafield>
+  <datafield tag="300" ind1=" " ind2=" ">
+    <subfield code="a">Online-Ressource</subfield>
+  </datafield>
+  <datafield tag="336" ind1=" " ind2=" ">
+    <subfield code="b">txt</subfield>
+  </datafield>
+  <datafield tag="337" ind1=" " ind2=" ">
+    <subfield code="b">c</subfield>
+  </datafield>
+  <datafield tag="338" ind1=" " ind2=" ">
+    <subfield code="b">cr</subfield>
+  </datafield>
+  <datafield tag="856" ind1="4" ind2=" ">
+    <subfield code="u">http://www.bibliothek.uni-regensburg.de/ezeit/?2102077</subfield>
+    <subfield code="x">EZB</subfield>
+  </datafield>
+  <datafield tag="776" ind1="0" ind2="8">
+    <subfield code="i">Erscheint auch als</subfield>
+    <subfield code="n">Druck-Ausgabe</subfield>
+    <subfield code="t">Know your world extra</subfield>
+    <subfield code="x">0163-4844</subfield>
+    <subfield code="0">http://worldcat.org/issn/0163-4844</subfield>
+  </datafield>
+  <datafield tag="655" ind1=" " ind2="7">
+    <subfield code="a">Zeitschrift</subfield>
+    <subfield code="0">(DE-588)4067488-5</subfield>
+    <subfield code="2">gnd-content</subfield>
+    <subfield code="B">GND-040674886</subfield>
+  </datafield>
+  <datafield tag="MBD" ind1=" " ind2=" ">
+    <subfield code="M">49HBZ_NETWORK</subfield>
+    <subfield code="i">990119186660206441</subfield>
+    <subfield code="n">HBZ Network</subfield>
+  </datafield>
+  <datafield tag="MNG" ind1=" " ind2=" ">
+    <subfield code="c">System</subfield>
+    <subfield code="f">ZDB</subfield>
+    <subfield code="i">marc21</subfield>
+    <subfield code="h">20230525135732.0</subfield>
+    <subfield code="k">01</subfield>
+    <subfield code="e">false</subfield>
+    <subfield code="d">2023-05-25 14:20:35 Europe/Berlin</subfield>
+    <subfield code="g">011918666-HBZ01</subfield>
+    <subfield code="j">90</subfield>
+    <subfield code="a">import</subfield>
+    <subfield code="b">2021-04-05 09:58:24 Europe/Berlin</subfield>
+  </datafield>
+  <datafield tag="GST" ind1=" " ind2=" ">
+    <subfield code="a">Periodikum</subfield>
+    <subfield code="A">GND-CONTENT</subfield>
+    <subfield code="B">GND-040674886</subfield>
+    <subfield code="C">450</subfield>
+  </datafield>
+  <datafield tag="GST" ind1=" " ind2=" ">
+    <subfield code="a">Zeitschriften</subfield>
+    <subfield code="A">GND-CONTENT</subfield>
+    <subfield code="B">GND-040674886</subfield>
+    <subfield code="C">450</subfield>
+  </datafield>
+  <datafield tag="GSI" ind1="7" ind2=" ">
+    <subfield code="a">4067488-5</subfield>
+    <subfield code="0">http://d-nb.info/gnd/4067488-5</subfield>
+    <subfield code="2">gnd</subfield>
+    <subfield code="A">GND-CONTENT</subfield>
+    <subfield code="B">GND-040674886</subfield>
+    <subfield code="C">024</subfield>
+  </datafield>
+</record>
diff --git a/web/test/tests/IndexIntegrationTest.java b/web/test/tests/IndexIntegrationTest.java
@@ -66,7 +66,7 @@ public static Collection<Object[]> data() {
 			{ "publication.startDate:1993", /*->*/ 3 },
 			{ "publication.location:Berlin AND publication.startDate:1993", /*->*/ 1 },
 			{ "publication.location:Berlin AND publication.startDate:[1992 TO 2017]", /*->*/ 4 },
-			{ "inCollection.id:\"http\\://lobid.org/organisations/DE-655#\\!\"", /*->*/ 129 },
+			{ "inCollection.id:\"http\\://lobid.org/organisations/DE-655#\\!\"", /*->*/ 130 },
 			{ "inCollection.id:NWBib", /*->*/ 0 },
 			{ "publication.publishedBy:Quedenfeldt", /*->*/ 2 },
 			{ "publication.publishedBy:Quedenfeld", /*->*/ 2 },