diff --git a/tabulate/__init__.py b/tabulate/__init__.py
index 12a2950..a743000 100644
--- a/tabulate/__init__.py
+++ b/tabulate/__init__.py
@@ -2504,7 +2504,7 @@ def _pad_row(cells, padding):
         return cells
 
 
-def _build_simple_row(padded_cells: list[list], rowfmt: DataRow) -> str:
+def _build_simple_row(padded_cells: list[list], rowfmt: DataRow, preserve_trailing_empty: bool = False) -> str:
     "Format row according to DataRow format without padding."
     begin = rowfmt.begin
     sep = rowfmt.sep
@@ -2520,7 +2520,15 @@ def escape_char(c):
     else:
         escaped_cells = padded_cells
 
-    return (begin + sep.join(escaped_cells) + end).rstrip()
+    row = begin + sep.join(escaped_cells) + end
+    # When there is no closing delimiter and the last cell is purely whitespace
+    # (i.e. a missing/None value formatted as ""), rstrip() silently drops the
+    # trailing column separator and makes that column invisible.  Skip rstrip()
+    # only for primary data rows (preserve_trailing_empty=True) so that
+    # multiline continuation lines are still stripped as before.
+    if preserve_trailing_empty and not end and escaped_cells and not escaped_cells[-1].strip():
+        return row
+    return row.rstrip()
 
 
 def _build_row(
@@ -2528,6 +2536,7 @@ def _build_row(
     colwidths: list[int],
     colaligns: list[str],
     rowfmt: DataRow | Callable,
+    preserve_trailing_empty: bool = False,
 ) -> str:
     "Return a string which represents a row of data cells."
     if not rowfmt:
@@ -2535,12 +2544,12 @@ def _build_row(
     if callable(rowfmt):
         return rowfmt(padded_cells, colwidths, colaligns)
     else:
-        return _build_simple_row(padded_cells, rowfmt)
+        return _build_simple_row(padded_cells, rowfmt, preserve_trailing_empty)
 
 
-def _append_basic_row(lines, padded_cells, colwidths, colaligns, rowfmt, rowalign=None):
+def _append_basic_row(lines, padded_cells, colwidths, colaligns, rowfmt, rowalign=None, preserve_trailing_empty=False):
     # NOTE: rowalign is ignored and exists for api compatibility with _append_multiline_row
-    lines.append(_build_row(padded_cells, colwidths, colaligns, rowfmt))
+    lines.append(_build_row(padded_cells, colwidths, colaligns, rowfmt, preserve_trailing_empty))
     return lines
 
 
@@ -2624,7 +2633,9 @@ def _format_table(
         append_row = partial(_append_multiline_row, pad=pad)
     else:
         pad_row = _pad_row
-        append_row = _append_basic_row
+        # preserve_trailing_empty=True so that a trailing None/empty cell (which
+        # formats to "") is not silently rstripped away and the column disappears.
+        append_row = partial(_append_basic_row, preserve_trailing_empty=True)
 
     padded_headers = pad_row(headers, pad)
 
diff --git a/test/test_input.py b/test/test_input.py
index 3cc3237..e55e8fe 100644
--- a/test/test_input.py
+++ b/test/test_input.py
@@ -57,7 +57,7 @@ def test_list_of_lists():
             "    string      number",
             "--  --------  --------",
             "a   one              1",
-            "b   two",
+            "b   two               ",
         ]
     )
     result = tabulate(ll, headers=["string", "number"])
@@ -72,7 +72,7 @@ def test_list_of_lists_firstrow():
             "    string      number",
             "--  --------  --------",
             "a   one              1",
-            "b   two",
+            "b   two               ",
         ]
     )
     result = tabulate(ll, headers="firstrow")
@@ -82,7 +82,7 @@ def test_list_of_lists_firstrow():
 def test_list_of_lists_keys():
     "Input: a list of lists with column indices as headers."
     ll = [["a", "one", 1], ["b", "two", None]]
-    expected = "\n".join(["0    1      2", "---  ---  ---", "a    one    1", "b    two"])
+    expected = "\n".join(["0    1      2", "---  ---  ---", "a    one    1", "b    two     "])
     result = tabulate(ll, headers="keys")
     assert_equal(expected, result)
 
@@ -405,8 +405,8 @@ def test_list_of_dicts_with_missing_keys():
         [
             "  foo    bar    baz",
             "-----  -----  -----",
-            "    1",
-            "           2",
+            "    1              ",
+            "           2       ",
             "    4             3",
         ]
     )
diff --git a/test/test_output.py b/test/test_output.py
index ea3da87..8f6f333 100644
--- a/test/test_output.py
+++ b/test/test_output.py
@@ -3028,7 +3028,7 @@ def test_missingval_multi():
         missingval=("n/a", "?"),
         tablefmt="plain",
     )
-    expected = "Alice  Bob  Charlie\nn/a    ?"
+    expected = "Alice  Bob  Charlie\nn/a    ?           "
     assert_equal(expected, result)
 
 
@@ -3053,7 +3053,7 @@ def test_column_emptymissing_deduction():
     ?              1.2342  1/3
     0.056789  12,345       abc
                            ?
-3,333.3            ?
+3,333.3            ?          
 ------------  -----------  ---"""
     assert_equal(expected, result)
 
diff --git a/test/test_regression.py b/test/test_regression.py
index 9555676..1ed2ef6 100644
--- a/test/test_regression.py
+++ b/test/test_regression.py
@@ -205,7 +205,7 @@ def test_88_256_ANSI_color_codes():
 
 def test_column_with_mixed_value_types():
     "Regression: mixed value types in the same column (issue #31)"
-    expected = "\n".join(["-----", "", "a", "я", "0", "False", "-----"])
+    expected = "\n".join(["-----", "     ", "a", "я", "0", "False", "-----"])
     data = [[None], ["a"], ["\u044f"], [0], [False]]
     table = tabulate(data)
     assert_equal(table, expected)
@@ -415,7 +415,7 @@ def test_escape_empty_cell_in_first_column_in_rst():
 def test_ragged_rows():
     "Regression: allow rows with different number of columns (issue #85)"
     table = [[1, 2, 3], [1, 2], [1, 2, 3, 4]]
-    expected = "\n".join(["-  -  -  -", "1  2  3", "1  2", "1  2  3  4", "-  -  -  -"])
+    expected = "\n".join(["-  -  -  -", "1  2  3   ", "1  2      ", "1  2  3  4", "-  -  -  -"])
     result = tabulate(table)
     assert_equal(expected, result)