From ba40d91de67ff869f87c403d098c8500733560be Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Saugat=20Pachhai=20=28=E0=A4=B8=E0=A5=8C=E0=A4=97=E0=A4=BE?=
 =?UTF-8?q?=E0=A4=A4=29?= <suagatchhetri@outlook.com>
Date: Wed, 22 Dec 2021 19:54:21 +0545
Subject: [PATCH] fs.hdfs: skip port if not available

---
 dvc/fs/hdfs.py             |  6 +++---
 tests/unit/fs/test_hdfs.py | 16 ++++++++++++++++
 2 files changed, 19 insertions(+), 3 deletions(-)
 create mode 100644 tests/unit/fs/test_hdfs.py

diff --git a/dvc/fs/hdfs.py b/dvc/fs/hdfs.py
index 2f7b8f56d..e77fdf9dd 100644
--- a/dvc/fs/hdfs.py
+++ b/dvc/fs/hdfs.py
@@ -27,9 +27,9 @@ class HDFSFileSystem(CallbackMixin, FSSpecWrapper):
 
     def unstrip_protocol(self, path: str) -> str:
         host = self.fs_args["host"]
-        port = self.fs_args["port"]
-        path = path.lstrip("/")
-        return f"hdfs://{host}:{port}/{path}"
+        port = self.fs_args.get("port")
+        netloc = host + (f":{port}" if port else "")
+        return "hdfs://" + netloc + "/" + path.lstrip("/")
 
     @staticmethod
     def _get_kwargs_from_urls(urlpath):
diff --git a/tests/unit/fs/test_hdfs.py b/tests/unit/fs/test_hdfs.py
new file mode 100644
index 000000000..fee6b735e
--- /dev/null
+++ b/tests/unit/fs/test_hdfs.py
@@ -0,0 +1,16 @@
+import pytest
+
+from dvc.fs import HDFSFileSystem
+
+
+@pytest.mark.parametrize(
+    "fs_args, expected_url",
+    [
+        ({"host": "example.com"}, "hdfs://example.com"),
+        ({"host": "example.com", "port": None}, "hdfs://example.com"),
+        ({"host": "example.com", "port": 8020}, "hdfs://example.com:8020"),
+    ],
+)
+def test_hdfs_unstrip_protocol(fs_args, expected_url):
+    fs = HDFSFileSystem(**fs_args)
+    assert fs.unstrip_protocol("/path") == expected_url + "/path"