diff --git a/bench.py b/bench.py
index 0841030..cad0cac 100755
--- a/bench.py
+++ b/bench.py
@@ -18,6 +18,7 @@ def run1(args, src_name, num_runs):
 runs = [
 	[['java', '-jar', './bin/freq01scala.jar'], 'freq01.scala', 3],
 	[['python', './src/freq01.py'], 'freq01.py', 3],
+	[['python', './src/freq02.py'], 'freq02.py', 3],
 	[['./bin/freq03cpp' + EXE], 'freq03.cpp'],
 	[['./bin/freq02cpp' + EXE], 'freq02.cpp'],
 	[['./bin/freq01cpp' + EXE], 'freq01.cpp'],
diff --git a/src/freq02.py b/src/freq02.py
new file mode 100644
index 0000000..8b52091
--- /dev/null
+++ b/src/freq02.py
@@ -0,0 +1,69 @@
+import string
+import sys
+from collections import Counter
+
+from itertools import chain
+
+
+CHUNK_SIZE = 20 * 1024 * 1024
+
+
+non_alpha = bytes(ch for ch in range(256) if chr(ch) not in string.ascii_letters)
+
+translation_table = bytes.maketrans(
+    string.ascii_uppercase.encode() + non_alpha,
+    string.ascii_lowercase.encode() + b' ' * len(non_alpha)
+)
+
+tail = [b'']  # variable to store tail of previous chunk of data.
+
+
+def get_words_with_empty_strings(line):
+    """Replaces any non-alphabetical symbols with spaces and splits the string using space separator.
+
+    Saves the last word into tails[0] to concat it with the first word of the next chunk.
+    """
+    result = line.translate(translation_table).split(b' ')
+    result[0] = tail[0] + result[0]
+    tail[0] = result[-1]
+    result[-1] = b''
+    return result
+
+
+def sort_key(x):
+    word, fq = x
+    return -fq, word
+
+
+def main(in_file_name, out_file_name):
+
+    with open(in_file_name, 'rb') as in_file:
+        def read_chunk():
+            return in_file.read(CHUNK_SIZE)
+
+        words_and_spaces = chain.from_iterable(
+            get_words_with_empty_strings(line) for line in iter(read_chunk, b'')
+        )
+
+        # Ignore empty strings.
+        words = filter(None, chain(words_and_spaces, tail[0]))
+
+        counts = sorted(Counter(words).items(), key=sort_key)
+
+        rows = (
+            b' '.join((str(v).encode(), k))
+            for k, v in counts
+        )
+
+        result = b'\r\n'.join(rows) + b'\r\n'
+
+    with open(out_file_name, 'wb+') as out_file:
+        out_file.write(result)
+
+
+if __name__ == '__main__':
+    if len(sys.argv) != 3:
+        print('No args')
+        exit(1)
+
+    main(sys.argv[1], sys.argv[2])