diff --git a/clearml/datasets/dataset.py b/clearml/datasets/dataset.py
index b8ea6317..e6d91068 100644
--- a/clearml/datasets/dataset.py
+++ b/clearml/datasets/dataset.py
@@ -3639,6 +3639,8 @@ class Dataset(object):
         # noinspection PyBroadException
         try:
             if StorageManager.exists_file(source_url):
+                # handle local path provided without scheme
+                source_url = StorageHelper.sanitize_url(source_url)
                 remote_objects = [StorageManager.get_metadata(source_url, return_full_path=True)]
             elif not source_url.startswith(("http://", "https://")):
                 if source_url[-1] != "/":
@@ -3655,7 +3657,7 @@ class Dataset(object):
             link = remote_object.get("name")
             relative_path = link[len(source_url) :]
             if not relative_path:
-                relative_path = source_url.split("/")[-1]
+                relative_path = os.path.basename(source_url)
             if not matches_any_wildcard(relative_path, wildcard, recursive=recursive):
                 continue
             try:
diff --git a/clearml/storage/helper.py b/clearml/storage/helper.py
index e08173c8..9e0d234b 100644
--- a/clearml/storage/helper.py
+++ b/clearml/storage/helper.py
@@ -3507,6 +3507,14 @@ class StorageHelper(object):
             container_name=self._container.name if self._container else "",
             object_name=object_name,
         )
+    
+    @classmethod
+    def sanitize_url(cls, remote_url):
+        base_url = cls._resolve_base_url(remote_url)
+        if base_url != 'file://' or remote_url.startswith("file://"):
+            return remote_url
+        absoulte_path = os.path.abspath(remote_url)
+        return base_url + absoulte_path
 
 
 def normalize_local_path(local_path: str) -> Path: