ENH: raise a more a useful exception on empty files

Author: cpcloud

doc/source/release.rst

@@ -96,6 +96,8 @@ pandas 0.11.1
     explicitly checking a website as a proxy for seeing if there is network
     connectivity. Plus, new ``optional_args`` decorator factory for decorators.
     (:issue:`3910`, :issue:`3914`)
+  - ``read_csv`` will now throw a more informative error message when a file
+    contains no columns, e.g., all newline characters
 
 **API Changes**
 

doc/source/v0.11.1.txt

@@ -240,6 +240,9 @@ I/O Enhancements
          import os
          os.remove(path)
 
+  - ``read_csv`` will now throw a more informative error message when a file
+    contains no columns, e.g., all newline characters
+
 Other Enhancements
 ~~~~~~~~~~~~~~~~~~
 

pandas/io/parsers.py

@@ -949,6 +949,7 @@ def __init__(self, src, **kwds):
 
         # #2442
         kwds['allow_leading_cols'] = self.index_col is not False
+
         self._reader = _parser.TextReader(src, **kwds)
 
         # XXX

pandas/io/tests/test_parsers.py

@@ -2330,6 +2330,19 @@ def test_tokenize_CR_with_quoting(self):
         expected = self.read_csv(StringIO(data.replace('\r', '\n')))
         tm.assert_frame_equal(result, expected)
 
+    def test_raise_on_no_columns(self):
+        # single newline
+        data = """
+"""
+        self.assertRaises(ValueError, self.read_csv, StringIO(data))
+
+        # test with more than a single newline
+        data = """
+
+
+"""
+        self.assertRaises(ValueError, self.read_csv, StringIO(data))
+
 
 class TestParseSQL(unittest.TestCase):
 

pandas/parser.pyx

@@ -476,6 +476,9 @@ cdef class TextReader:
         self.names = names
         self.header, self.table_width = self._get_header()
 
+        if not self.table_width:
+            raise ValueError("No columns to parse from file")
+
         # compute buffer_lines as function of table width
         heuristic = 2**20 // self.table_width
         self.buffer_lines = 1