diff --git a/swh/dataset/exporter.py b/swh/dataset/exporter.py
--- a/swh/dataset/exporter.py
+++ b/swh/dataset/exporter.py
@@ -7,6 +7,7 @@
 import pathlib
 from types import TracebackType
 from typing import Any, Dict, Optional, Type
+import uuid
 
 
 class Exporter:
@@ -53,6 +54,14 @@
         """
         raise NotImplementedError
 
+    def get_unique_file_id(self) -> str:
+        """
+        Return a unique random file id for the current process.
+
+        If config['test_unique_file_id'] is set, it will be used instead.
+        """
+        return str(self.config.get("test_unique_file_id", uuid.uuid4()))
+
 
 class ExporterDispatch(Exporter):
     """
diff --git a/swh/dataset/exporters/edges.py b/swh/dataset/exporters/edges.py
--- a/swh/dataset/exporters/edges.py
+++ b/swh/dataset/exporters/edges.py
@@ -10,7 +10,6 @@
 import subprocess
 import tempfile
 from typing import Tuple
-import uuid
 
 from swh.dataset.exporter import ExporterDispatch
 from swh.dataset.utils import ZSTFile, remove_pull_requests
@@ -42,7 +41,7 @@
         if obj_type not in self.writers:
             dataset_path = self.export_path / obj_type.name.lower()
             dataset_path.mkdir(exist_ok=True)
-            unique_id = str(uuid.uuid4())
+            unique_id = self.get_unique_file_id()
             nodes_file = dataset_path / ("graph-{}.nodes.csv.zst".format(unique_id))
             edges_file = dataset_path / ("graph-{}.edges.csv.zst".format(unique_id))
             node_writer = self.exit_stack.enter_context(ZSTFile(str(nodes_file), "w"))
diff --git a/swh/dataset/exporters/orc.py b/swh/dataset/exporters/orc.py
--- a/swh/dataset/exporters/orc.py
+++ b/swh/dataset/exporters/orc.py
@@ -6,7 +6,6 @@
 from datetime import datetime
 import math
 from typing import Any, Optional, Tuple, Type, cast
-import uuid
 
 from pyorc import (
     BigInt,
@@ -117,7 +116,7 @@
         if table_name not in self.writers:
             object_type_dir = self.export_path / table_name
             object_type_dir.mkdir(exist_ok=True)
-            unique_id = str(uuid.uuid4())
+            unique_id = self.get_unique_file_id()
             export_file = object_type_dir / ("graph-{}.orc".format(unique_id))
             export_obj = self.exit_stack.enter_context(export_file.open("wb"))
             self.writers[table_name] = self.exit_stack.enter_context(
diff --git a/swh/dataset/utils.py b/swh/dataset/utils.py
--- a/swh/dataset/utils.py
+++ b/swh/dataset/utils.py
@@ -29,7 +29,7 @@
         is_text = not (self.mode in ("rb", "wb"))
         writing = self.mode in ("w", "wb")
         if writing:
-            cmd = ["zstd", "-q", "-o", self.path]
+            cmd = ["zstd", "-f", "-q", "-o", self.path]
         else:
             cmd = ["zstdcat", self.path]
         self.process = subprocess.Popen(