diff --git a/test/s3/parquet/s3_parquet_test.py b/test/s3/parquet/s3_parquet_test.py
index 221116b75..1c8403e85 100755
--- a/test/s3/parquet/s3_parquet_test.py
+++ b/test/s3/parquet/s3_parquet_test.py
@@ -173,6 +173,16 @@ def read_with_pads_dataset(path: str, fs: s3fs.S3FileSystem) -> Tuple[bool, str,
 def read_direct_s3fs(path: str, fs: s3fs.S3FileSystem) -> Tuple[bool, str, int]:
     """Read directly via s3fs.open() streaming."""
     try:
+        # Check if path is a directory (common with pads.write_dataset)
+        if fs.isdir(path):
+            # Find parquet files in the directory
+            files = [f for f in fs.ls(path) if f.endswith('.parquet')]
+            if not files:
+                raise ValueError(f"No parquet files found in directory: {path}")
+            # For simplicity, read the first parquet file
+            # In production, you might want to read all files and concatenate
+            path = files[0]
+        
         with fs.open(path, "rb") as f:
             result = pq.read_table(f)
         return True, "s3fs.open+pq.read_table", result.num_rows
@@ -185,6 +195,16 @@ def read_direct_s3fs(path: str, fs: s3fs.S3FileSystem) -> Tuple[bool, str, int]:
 def read_buffered_s3fs(path: str, fs: s3fs.S3FileSystem) -> Tuple[bool, str, int]:
     """Read via s3fs.open() into buffer, then pq.read_table."""
     try:
+        # Check if path is a directory (common with pads.write_dataset)
+        if fs.isdir(path):
+            # Find parquet files in the directory
+            files = [f for f in fs.ls(path) if f.endswith('.parquet')]
+            if not files:
+                raise ValueError(f"No parquet files found in directory: {path}")
+            # For simplicity, read the first parquet file
+            # In production, you might want to read all files and concatenate
+            path = files[0]
+        
         with fs.open(path, "rb") as f:
             buffer = io.BytesIO(f.read())
         buffer.seek(0)