BUG (output formatting): use fixed with for truncation column instead of inferring from last column (#24905)

Author: jorisvandenbossche

doc/source/whatsnew/v0.24.0.rst

@@ -1755,6 +1755,7 @@ I/O
 - Bug in :meth:`DataFrame.to_stata` and :class:`pandas.io.stata.StataWriter117` that produced invalid files when using strLs with non-ASCII characters (:issue:`23573`)
 - Bug in :class:`HDFStore` that caused it to raise ``ValueError`` when reading a Dataframe in Python 3 from fixed format written in Python 2 (:issue:`24510`)
 - Bug in :func:`DataFrame.to_string()` and more generally in the floating ``repr`` formatter. Zeros were not trimmed if ``inf`` was present in a columns while it was the case with NA values. Zeros are now trimmed as in the presence of NA (:issue:`24861`).
+- Bug in the ``repr`` when truncating the number of columns and having a wide last column (:issue:`24849`).
 
 Plotting
 ^^^^^^^^

pandas/io/formats/format.py

@@ -435,9 +435,6 @@ def _chk_truncate(self):
         """
         from pandas.core.reshape.concat import concat
 
-        # Column of which first element is used to determine width of a dot col
-        self.tr_size_col = -1
-
         # Cut the data to the information actually printed
         max_cols = self.max_cols
         max_rows = self.max_rows
@@ -556,10 +553,7 @@ def _to_str_columns(self):
 
         if truncate_h:
             col_num = self.tr_col_num
-            # infer from column header
-            col_width = self.adj.len(strcols[self.tr_size_col][0])
-            strcols.insert(self.tr_col_num + 1, ['...'.center(col_width)] *
-                           (len(str_index)))
+            strcols.insert(self.tr_col_num + 1, [' ...'] * (len(str_index)))
         if truncate_v:
             n_header_rows = len(str_index) - len(frame)
             row_num = self.tr_row_num
@@ -577,8 +571,8 @@ def _to_str_columns(self):
                 if ix == 0:
                     dot_mode = 'left'
                 elif is_dot_col:
-                    cwidth = self.adj.len(strcols[self.tr_size_col][0])
-                    dot_mode = 'center'
+                    cwidth = 4
+                    dot_mode = 'right'
                 else:
                     dot_mode = 'right'
                 dot_str = self.adj.justify([my_str], cwidth, mode=dot_mode)[0]

pandas/tests/io/formats/test_format.py

@@ -345,6 +345,15 @@ def test_repr_truncates_terminal_size_full(self, monkeypatch):
                             lambda: terminal_size)
         assert "..." not in str(df)
 
+    def test_repr_truncation_column_size(self):
+        # dataframe with last column very wide -> check it is not used to
+        # determine size of truncation (...) column
+        df = pd.DataFrame({'a': [108480, 30830], 'b': [12345, 12345],
+                           'c': [12345, 12345], 'd': [12345, 12345],
+                           'e': ['a' * 50] * 2})
+        assert "..." in str(df)
+        assert "    ...    " not in str(df)
+
     def test_repr_max_columns_max_rows(self):
         term_width, term_height = get_terminal_size()
         if term_width < 10 or term_height < 10:
@@ -543,7 +552,7 @@ def test_to_string_with_formatters_unicode(self):
             formatters={u('c/\u03c3'): lambda x: '{x}'.format(x=x)})
         assert result == u('  c/\u03c3\n') + '0   1\n1   2\n2   3'
 
-    def test_east_asian_unicode_frame(self):
+    def test_east_asian_unicode_false(self):
         if PY3:
             _rep = repr
         else:
@@ -643,17 +652,23 @@ def test_east_asian_unicode_frame(self):
                                u'ああああ': [u'さ', u'し', u'す', u'せ']},
                               columns=['a', 'b', 'c', u'ああああ'])
 
-            expected = (u"        a ...  ああああ\n0   あああああ ...     さ\n"
-                        u"..    ... ...   ...\n3     えええ ...     せ\n"
+            expected = (u"        a  ... ああああ\n0   あああああ  ...    さ\n"
+                        u"..    ...  ...  ...\n3     えええ  ...    せ\n"
                         u"\n[4 rows x 4 columns]")
             assert _rep(df) == expected
 
             df.index = [u'あああ', u'いいいい', u'う', 'aaa']
-            expected = (u"         a ...  ああああ\nあああ  あああああ ...     さ\n"
-                        u"..     ... ...   ...\naaa    えええ ...     せ\n"
+            expected = (u"         a  ... ああああ\nあああ  あああああ  ...    さ\n"
+                        u"..     ...  ...  ...\naaa    えええ  ...    せ\n"
                         u"\n[4 rows x 4 columns]")
             assert _rep(df) == expected
 
+    def test_east_asian_unicode_true(self):
+        if PY3:
+            _rep = repr
+        else:
+            _rep = unicode  # noqa
+
         # Emable Unicode option -----------------------------------------
         with option_context('display.unicode.east_asian_width', True):
 
@@ -757,18 +772,18 @@ def test_east_asian_unicode_frame(self):
                                    u'ああああ': [u'さ', u'し', u'す', u'せ']},
                                   columns=['a', 'b', 'c', u'ああああ'])
 
-                expected = (u"             a   ...    ああああ\n"
-                            u"0   あああああ   ...          さ\n"
-                            u"..         ...   ...         ...\n"
-                            u"3       えええ   ...          せ\n"
+                expected = (u"             a  ... ああああ\n"
+                            u"0   あああああ  ...       さ\n"
+                            u"..         ...  ...      ...\n"
+                            u"3       えええ  ...       せ\n"
                             u"\n[4 rows x 4 columns]")
                 assert _rep(df) == expected
 
                 df.index = [u'あああ', u'いいいい', u'う', 'aaa']
-                expected = (u"                 a   ...    ああああ\n"
-                            u"あああ  あああああ   ...          さ\n"
-                            u"...            ...   ...         ...\n"
-                            u"aaa         えええ   ...          せ\n"
+                expected = (u"                 a  ... ああああ\n"
+                            u"あああ  あああああ  ...       さ\n"
+                            u"...            ...  ...      ...\n"
+                            u"aaa         えええ  ...       せ\n"
                             u"\n[4 rows x 4 columns]")
                 assert _rep(df) == expected