Bulk loader: avoids to write duplicated values

lib/src/storage/mod.rs

@@ -1500,9 +1500,19 @@ impl FileBulkLoader {
     ) -> Result<PathBuf, StorageError> {
         let mut values = values.collect::<Vec<_>>();
         values.sort_unstable();
+        let deduplicated_values = values.iter().enumerate().filter_map(|(i, value)| {
+            if values
+                .get(i + 1)
+                .map_or(true, |next_value| value != next_value)
+            {
+                Some(value)
+            } else {
+                None
+            }
+        });
         let mut sst = self.storage.db.new_sst_file()?;
-        for t in values {
-            sst.insert_empty(&t)?;
+        for value in deduplicated_values {
+            sst.insert_empty(value)?;
         }
         sst.finish()
     }

lib/tests/store.rs

@@ -165,11 +165,12 @@ fn test_bulk_load_dataset() -> Result<(), Box<dyn Error>> {
     store
         .bulk_loader()
         .load_dataset(Cursor::new(GRAPH_DATA), DatasetFormat::TriG, None)?;
-    for q in quads(NamedNodeRef::new_unchecked(
-        "http://www.wikidata.org/wiki/Special:EntityData/Q90",
-    )) {
+    let graph_name =
+        NamedNodeRef::new_unchecked("http://www.wikidata.org/wiki/Special:EntityData/Q90");
+    for q in quads(graph_name) {
         assert!(store.contains(q)?);
     }
+    assert!(store.contains_named_graph(graph_name)?);
     store.validate()?;
     Ok(())
 }