diff --git a/python/cudf/cudf/_lib/text.pyx b/python/cudf/cudf/_lib/text.pyx
index daea227cc39..868574be187 100644
--- a/python/cudf/cudf/_lib/text.pyx
+++ b/python/cudf/cudf/_lib/text.pyx
@@ -1,5 +1,7 @@
 # Copyright (c) 2020-2022, NVIDIA CORPORATION.
 
+from io import TextIOBase
+
 import cudf
 
 from cython.operator cimport dereference
@@ -28,30 +30,33 @@ def read_text(object filepaths_or_buffers,
     --------
     cudf.io.text.read_text
     """
-    cdef string filename = filepaths_or_buffers.encode()
     cdef string delim = delimiter.encode()
 
     cdef unique_ptr[data_chunk_source] datasource
     cdef unique_ptr[column] c_col
+
     cdef size_t c_byte_range_offset
     cdef size_t c_byte_range_size
     cdef byte_range_info c_byte_range
 
-    if (byte_range is not None):
+    if isinstance(filepaths_or_buffers, TextIOBase):
+        datasource = move(make_source(filepaths_or_buffers.read().encode()))
+    else:
+        datasource = move(make_source_from_file(filepaths_or_buffers.encode()))
+
+    if (byte_range is None):
+        with nogil:
+            c_col = move(multibyte_split(dereference(datasource), delim))
+    else:
         c_byte_range_offset = byte_range[0]
         c_byte_range_size = byte_range[1]
+        c_byte_range = byte_range_info(
+            c_byte_range_offset,
+            c_byte_range_size)
         with nogil:
-            datasource = move(make_source_from_file(filename))
-            c_byte_range = byte_range_info(
-                c_byte_range_offset,
-                c_byte_range_size)
             c_col = move(multibyte_split(
                 dereference(datasource),
                 delim,
                 c_byte_range))
-    else:
-        with nogil:
-            datasource = move(make_source_from_file(filename))
-            c_col = move(multibyte_split(dereference(datasource), delim))
 
     return {None: Column.from_unique_ptr(move(c_col))}
diff --git a/python/cudf/cudf/tests/test_text.py b/python/cudf/cudf/tests/test_text.py
index fb6505f5f92..c332924fd8b 100644
--- a/python/cudf/cudf/tests/test_text.py
+++ b/python/cudf/cudf/tests/test_text.py
@@ -1,5 +1,7 @@
 # Copyright (c) 2019-2022, NVIDIA CORPORATION.
 
+from io import StringIO
+
 import numpy as np
 import pytest
 
@@ -829,3 +831,13 @@ def test_read_text_byte_range_large(datadir):
         f.write(content)
 
     cudf.read_text(temp_file, delimiter=delimiter)
+
+
+def test_read_text_in_memory(datadir):
+    # Since Python split removes the delimiter and read_text does
+    # not we need to add it back to the 'content'
+    expected = cudf.Series(["x::", "y::", "z"])
+
+    actual = cudf.read_text(StringIO("x::y::z"), delimiter="::")
+
+    assert_eq(expected, actual)