test read_csv with header=None

mrocklin · Apr 30, 2016 · 6a477d9 · 6a477d9
1 parent 864492f
commit 6a477d9
Show file tree

Hide file tree

Showing 2 changed files with 13 additions and 1 deletion.
diff --git a/dask/dataframe/csv.py b/dask/dataframe/csv.py
@@ -175,7 +175,10 @@ def read_csv(filename, blocksize=2**25, chunkbytes=None,
     if 'nrows' in kwargs:
         values = [[values[0][0]]]
 
-    header = sample.split(b_lineterminator)[0] + b_lineterminator
+    if kwargs.get('header', 'infer') is None:
+        header = b''
+    else:
+        header = sample.split(b_lineterminator)[0] + b_lineterminator
     head = pd.read_csv(BytesIO(sample), **kwargs)
 
     df = read_csv_from_bytes(values, header, head, kwargs,

diff --git a/dask/dataframe/tests/test_csv.py b/dask/dataframe/tests/test_csv.py
@@ -187,3 +187,12 @@ def test_late_dtypes():
 
         assert df.a.sum().compute() == 1 + 2 + 3 + 4 + 5.5 + 6
         assert df.b.sum().compute() == 2 + 3 + 4 + 5 + 6 + 7.5
+
+
+def test_header_None():
+    with filetexts({'.tmp.1.csv': '1,2',
+                    '.tmp.2.csv': '',
+                    '.tmp.3.csv': '3,4'}):
+        df = read_csv('.tmp.*.csv', header=None)
+        expected = pd.DataFrame({0: [1, 3], 1: [2, 4]})
+        eq(df.compute().reset_index(drop=True), expected)