Bug Fix: Return the parsed path without netloc for HDFS (#675)

Author: frankliee

pyiceberg/io/pyarrow.py

@@ -333,7 +333,7 @@ def parse_location(location: str) -> Tuple[str, str, str]:
         if not uri.scheme:
             return "file", uri.netloc, os.path.abspath(location)
         elif uri.scheme == "hdfs":
-            return uri.scheme, uri.netloc, location
+            return uri.scheme, uri.netloc, uri.path
         else:
             return uri.scheme, uri.netloc, f"{uri.netloc}{uri.path}"
 

tests/io/test_pyarrow.py

@@ -1644,9 +1644,9 @@ def check_results(location: str, expected_schema: str, expected_netloc: str, exp
         assert netloc == expected_netloc
         assert uri == expected_uri
 
-    check_results("hdfs://127.0.0.1:9000/root/foo.txt", "hdfs", "127.0.0.1:9000", "hdfs://127.0.0.1:9000/root/foo.txt")
-    check_results("hdfs://127.0.0.1/root/foo.txt", "hdfs", "127.0.0.1", "hdfs://127.0.0.1/root/foo.txt")
-    check_results("hdfs://clusterA/root/foo.txt", "hdfs", "clusterA", "hdfs://clusterA/root/foo.txt")
+    check_results("hdfs://127.0.0.1:9000/root/foo.txt", "hdfs", "127.0.0.1:9000", "/root/foo.txt")
+    check_results("hdfs://127.0.0.1/root/foo.txt", "hdfs", "127.0.0.1", "/root/foo.txt")
+    check_results("hdfs://clusterA/root/foo.txt", "hdfs", "clusterA", "/root/foo.txt")
 
     check_results("/root/foo.txt", "file", "", "/root/foo.txt")
     check_results("/root/tmp/foo.txt", "file", "", "/root/tmp/foo.txt")