BUG: DataFrame column formatting issue in length-truncated column close pandas-dev#1906

Author: wesm

RELEASE.rst

@@ -160,6 +160,7 @@ pandas 0.9.0
     window (#1884)
   - Fix unicode sheet name failure in to_excel (#1828)
   - Override DatetimeIndex.min/max to return Timestamp objects (#1895)
+  - Fix column name formatting issue in length-truncated column (#1906)
 
 pandas 0.8.1
 ============

pandas/core/format.py

@@ -219,15 +219,20 @@ def _to_str_columns(self, force_unicode=False):
             if self.header:
                 fmt_values = self._format_col(i)
                 cheader = str_columns[i]
+
+                max_colwidth = max(_strlen(x) for x in cheader)
+
+                fmt_values = _make_fixed_width(fmt_values, self.justify,
+                                               minimum=max_colwidth)
+
                 max_len = max(max(_strlen(x) for x in fmt_values),
-                              max(len(x) for x in cheader))
+                              max_colwidth)
                 if self.justify == 'left':
                     cheader = [x.ljust(max_len) for x in cheader]
                 else:
                     cheader = [x.rjust(max_len) for x in cheader]
-                fmt_values = cheader + fmt_values
-                stringified.append(_make_fixed_width(fmt_values,
-                                                     self.justify))
+
+                stringified.append(cheader + fmt_values)
             else:
                 stringified = [_make_fixed_width(self._format_col(i),
                                                  self.justify)
@@ -704,11 +709,15 @@ def _format_datetime64(x, tz=None):
     return stamp._repr_base
 
 
-def _make_fixed_width(strings, justify='right'):
+def _make_fixed_width(strings, justify='right', minimum=None):
     if len(strings) == 0:
         return strings
 
     max_len = max(_strlen(x) for x in strings)
+
+    if minimum is not None:
+        max_len = max(minimum, max_len)
+
     conf_max = print_config.max_colwidth
     if conf_max is not None and max_len > conf_max:
         max_len = conf_max

pandas/tests/test_frame.py

@@ -2614,6 +2614,21 @@ def test_very_wide_info_repr(self):
                        columns=[tm.rands(10) for _ in xrange(20)])
         repr(df)
 
+    def test_repr_column_name_unicode_truncation_bug(self):
+        # #1906
+        df = DataFrame({'Id': [7117434],
+                        'StringCol': ('Is it possible to modify drop plot code'
+                                      ' so that the output graph is displayed '
+                                      'in iphone simulator, Is it possible to '
+                                      'modify drop plot code so that the '
+                                      'output graph is \xe2\x80\xa8displayed '
+                                      'in iphone simulator.Now we are adding '
+                                      'the CSV file externally. I want to Call'
+                                      ' the File through the code..')})
+
+        result = repr(df)
+        self.assert_('StringCol' in result)
+
     def test_head_tail(self):
         assert_frame_equal(self.frame.head(), self.frame[:5])
         assert_frame_equal(self.frame.tail(), self.frame[-5:])

pandas/tseries/tests/test_timeseries.py

@@ -2151,3 +2151,4 @@ def test_hash_equivalent(self):
 if __name__ == '__main__':
     nose.runmodule(argv=[__file__,'-vvs','-x','--pdb', '--pdb-failure'],
                    exit=False)
+