fix symlink display on marquez

CHANGELOG.md

@@ -16,6 +16,11 @@
 * Web: minor UI enhancements [`#2727`](https://github.com/MarquezProject/marquez/pull/2727) [@phixMe](https://github.com/phixMe)  
     *Hygienic cleanup of project as a follow-up to [`#2725`](https://github.com/MarquezProject/marquez/pull/2725), including a fix for [`#2747`](https://github.com/MarquezProject/marquez/issues/2747).*
 
+### Fixed
+
+* bug: marquez dataset symlinks facet create empty namespace: [`#2645`](https://github.com/MarquezProject/marquez/pull/2645) [@sophiely](https://github.com/sophiely)  
+Display symlink dataset in the previously empty namespace and link the symlink dataset lineage to the main dataset.
+
 ## [0.44.0](https://github.com/MarquezProject/marquez/compare/0.43.1...0.44.0) - 2024-01-22
 
 ### Added

api/src/main/java/marquez/db/LineageDao.java

@@ -102,16 +102,21 @@ INNER JOIN jobs_view j ON (j.uuid=l2.job_uuid OR j.uuid=l2.job_symlink_target_uu
       """
       SELECT ds.*, dv.fields, dv.lifecycle_state
       FROM datasets_view ds
-      LEFT JOIN dataset_versions dv on dv.uuid = ds.current_version_uuid
-      WHERE ds.uuid IN (<dsUuids>)""")
+      LEFT JOIN dataset_versions dv ON dv.uuid = ds.current_version_uuid
+      LEFT JOIN dataset_symlinks dsym ON dsym.namespace_uuid = ds.namespace_uuid and dsym.name = ds.name
+      WHERE dsym.is_primary = true
+      AND ds.uuid IN (<dsUuids>)""")
   Set<DatasetData> getDatasetData(@BindList Set<UUID> dsUuids);
 
   @SqlQuery(
       """
       SELECT ds.*, dv.fields, dv.lifecycle_state
       FROM datasets_view ds
       LEFT JOIN dataset_versions dv on dv.uuid = ds.current_version_uuid
-      WHERE ds.name = :datasetName AND ds.namespace_name = :namespaceName""")
+      LEFT JOIN dataset_symlinks dsym ON dsym.namespace_uuid = ds.namespace_uuid and dsym.name = ds.name
+      INNER JOIN datasets_view AS d ON d.uuid = ds.uuid
+      WHERE dsym.is_primary is true
+      AND CAST((:namespaceName, :datasetName) AS DATASET_NAME) = ANY(d.dataset_symlinks)""")
   DatasetData getDatasetData(String namespaceName, String datasetName);
 
   @SqlQuery(

api/src/main/java/marquez/service/LineageService.java

@@ -114,15 +114,20 @@ public Lineage lineage(NodeId nodeId, int depth, boolean withRunFacets) {
     if (!datasetIds.isEmpty()) {
       datasets.addAll(this.getDatasetData(datasetIds));
     }
-    if (nodeId.isDatasetType()
-        && datasets.stream().noneMatch(n -> n.getId().equals(nodeId.asDatasetId()))) {
-      log.warn(
-          "Found jobs {} which no longer share lineage with dataset '{}' - discarding",
-          jobData.stream().map(JobData::getId).toList(),
-          nodeId.getValue());
-      return toLineageWithOrphanDataset(nodeId.asDatasetId());
-    }
 
+    if (nodeId.isDatasetType()) {
+      DatasetId datasetId = nodeId.asDatasetId();
+      DatasetData datasetData =
+          this.getDatasetData(datasetId.getNamespace().getValue(), datasetId.getName().getValue());
+
+      if (!datasetIds.contains(datasetData.getUuid())) {
+        log.warn(
+            "Found jobs {} which no longer share lineage with dataset '{}' - discarding",
+            jobData.stream().map(JobData::getId).toList(),
+            nodeId.getValue());
+        return toLineageWithOrphanDataset(nodeId.asDatasetId());
+      }
+    }
     return toLineage(jobData, datasets);
   }
 

api/src/main/resources/marquez/db/migration/R__3_Datasets_view.sql

@@ -1,23 +1,33 @@
 DROP VIEW IF EXISTS datasets_view;
-CREATE VIEW datasets_view
-AS
+CREATE VIEW datasets_view AS
 SELECT d.uuid,
-       d.type,
-       d.created_at,
-       d.updated_at,
-       d.namespace_uuid,
-       d.source_uuid,
-       d.name,
-       array_agg(CAST((namespaces.name, symlinks.name) AS DATASET_NAME)) AS dataset_symlinks,
-       d.physical_name,
-       d.description,
-       d.current_version_uuid,
-       d.last_modified_at,
-       d.namespace_name,
-       d.source_name,
-       d.is_deleted
-FROM datasets d
-JOIN dataset_symlinks symlinks ON d.uuid = symlinks.dataset_uuid
-INNER JOIN namespaces ON symlinks.namespace_uuid = namespaces.uuid
-WHERE d.is_hidden IS FALSE
-GROUP BY d.uuid;
+    d.type,
+    d.created_at,
+    d.updated_at,
+    CASE
+    	 WHEN (d.namespace_name = namespaces.name AND d.name = symlinks.name) THEN d.namespace_uuid
+    	 ELSE namespaces.uuid
+    END
+    AS namespace_uuid ,
+    d.source_uuid,
+    CASE
+    	 WHEN (d.namespace_name = namespaces.name and d.name = symlinks.name) THEN d.name
+    	 ELSE symlinks.name
+    END
+    AS name,
+    array(SELECT ROW(namespaces.name::character varying(255), symlinks.name::character varying(255))::dataset_name) AS dataset_symlinks,
+    d.physical_name,
+    d.description,
+    d.current_version_uuid,
+    d.last_modified_at,
+    CASE
+    	 WHEN (d.namespace_name = namespaces.name AND d.name = symlinks.name) THEN d.namespace_name
+    	 ELSE namespaces.name
+    END
+    AS namespace_name,
+    d.source_name,
+    d.is_deleted
+   FROM datasets d
+     JOIN dataset_symlinks symlinks ON d.uuid = symlinks.dataset_uuid
+     JOIN namespaces ON symlinks.namespace_uuid = namespaces.uuid
+  WHERE d.is_hidden is false;

api/src/test/java/marquez/service/LineageServiceTest.java

@@ -13,8 +13,10 @@
 
 import java.util.Arrays;
 import java.util.Collections;
+import java.util.HashMap;
 import java.util.LinkedList;
 import java.util.List;
+import java.util.Map;
 import java.util.Optional;
 import java.util.UUID;
 import java.util.stream.Collectors;
@@ -588,4 +590,64 @@ public void testLineageForOrphanedDataset() {
   private boolean jobNameEquals(Node node, String writeJob) {
     return node.getId().asJobId().getName().getValue().equals(writeJob);
   }
+
+  @Test
+  public void testSymlinkDatasetLineage() {
+    // (1) Create symlink facet for our main dataset
+    Map<String, Object> symlink = new HashMap<>();
+    Map<String, Object> symlinkInfo = new HashMap<>();
+    Map<String, Object> symlinkIdentifiers = new HashMap<>();
+    symlinkIdentifiers.put("name", "symlinkDataset");
+    symlinkIdentifiers.put("namespace", NAMESPACE);
+    symlinkIdentifiers.put("type", "DB_TABLE");
+    symlinkInfo.put("producer", "https://github.com/OpenLineage/producer/");
+    symlinkInfo.put("schemaURL", "https://openlineage.io/schema/url/");
+    symlinkInfo.put("identifiers", symlinkIdentifiers);
+    symlink.put("symlinks", symlinkInfo);
+
+    // (2) Create main dataset with a symlink
+    Dataset mainDataset =
+        new Dataset(
+            NAMESPACE,
+            "mainDataset",
+            newDatasetFacet(symlink, new SchemaField("firstname", "string", "the first name")));
+
+    // (3) Create the symlink dataset
+    Dataset symlinkDataset =
+        new Dataset(
+            NAMESPACE,
+            "symlinkDataset",
+            newDatasetFacet(new SchemaField("firstname", "string", "the first name")));
+
+    // (3) Create a job with the main dataset
+    UpdateLineageRow firstJob =
+        LineageTestUtils.createLineageRow(
+            openLineageDao,
+            "firstJob",
+            "COMPLETE",
+            jobFacet,
+            Arrays.asList(mainDataset),
+            Arrays.asList());
+
+    // (4) Create a job with the symlink dataset
+    UpdateLineageRow secondJob =
+        LineageTestUtils.createLineageRow(
+            openLineageDao,
+            "secondJob",
+            "COMPLETE",
+            jobFacet,
+            Arrays.asList(symlinkDataset),
+            Arrays.asList());
+
+    // (5) We expect the first and second job linked together because the main
+    // and symlink dataset are in fact the same dataset
+    Lineage lineage =
+        lineageService.lineage(
+            NodeId.of(
+                new DatasetId(new NamespaceName(NAMESPACE), new DatasetName("symlinkDataset"))),
+            5,
+            true);
+
+    assertThat(lineage.getGraph()).hasSize(2);
+  }
 }