diff --git a/cpp/src/io/orc/reader_impl.cu b/cpp/src/io/orc/reader_impl.cu
index 61adef26dab..2567b2579d7 100644
--- a/cpp/src/io/orc/reader_impl.cu
+++ b/cpp/src/io/orc/reader_impl.cu
@@ -535,9 +535,7 @@ table_with_metadata reader::impl::read(size_type skip_rows,
           chunk.ts_clock_rate = to_clockrate(_timestamp_type.id());
         }
         for (int k = 0; k < gpu::CI_NUM_STREAMS; k++) {
-          if (chunk.strm_len[k] > 0) {
-            chunk.streams[k] = dst_base + stream_info[chunk.strm_id[k]].dst_pos;
-          }
+          chunk.streams[k] = dst_base + stream_info[chunk.strm_id[k]].dst_pos;
         }
       }
       stripe_start_row += stripe_info->numberOfRows;
diff --git a/python/cudf/cudf/tests/test_orc.py b/python/cudf/cudf/tests/test_orc.py
index ca8aa00f80c..fa14a0a9690 100644
--- a/python/cudf/cudf/tests/test_orc.py
+++ b/python/cudf/cudf/tests/test_orc.py
@@ -738,3 +738,19 @@ def test_nanoseconds_overflow():
 
     pyarrow_got = pa.orc.ORCFile(buffer).read()
     assert_eq(expected.to_pandas(), pyarrow_got.to_pandas())
+
+
+@pytest.mark.parametrize(
+    "data", [[None, ""], ["", None], [None, None], ["", ""]]
+)
+def test_empty_string_columns(data):
+    buffer = BytesIO()
+
+    expected = cudf.DataFrame({"string": data}, dtype="str")
+    expected.to_orc(buffer)
+
+    expected_pdf = pd.read_orc(buffer)
+    got_df = cudf.read_orc(buffer)
+
+    assert_eq(expected, got_df)
+    assert_eq(expected_pdf, got_df)