ER: multiindexes in no-data DataFrame constructor should yield a nan frame

[cpcloud]

as it stands

In [24]: tuples = [(randint(10), randint(10)) for _ in range(10)]

In [25]: mi = MultiIndex.from_tuples(tuples)

In [26]: df = DataFrame(index=mi, columns=mi)

fails, whereas single level indexes work fine

In [27]: df = DataFrame(index=range(10), columns=range(10))

In [28]: df
Out[28]:
     0    1    2    3    4    5    6    7    8    9
0  NaN  NaN  NaN  NaN  NaN  NaN  NaN  NaN  NaN  NaN
1  NaN  NaN  NaN  NaN  NaN  NaN  NaN  NaN  NaN  NaN
2  NaN  NaN  NaN  NaN  NaN  NaN  NaN  NaN  NaN  NaN
3  NaN  NaN  NaN  NaN  NaN  NaN  NaN  NaN  NaN  NaN
4  NaN  NaN  NaN  NaN  NaN  NaN  NaN  NaN  NaN  NaN
5  NaN  NaN  NaN  NaN  NaN  NaN  NaN  NaN  NaN  NaN
6  NaN  NaN  NaN  NaN  NaN  NaN  NaN  NaN  NaN  NaN
7  NaN  NaN  NaN  NaN  NaN  NaN  NaN  NaN  NaN  NaN
8  NaN  NaN  NaN  NaN  NaN  NaN  NaN  NaN  NaN  NaN
9  NaN  NaN  NaN  NaN  NaN  NaN  NaN  NaN  NaN  NaN

[jtratner]

Or at least a non-crazy error:

---------------------------------------------------------------------------
Exception                                 Traceback (most recent call last)
<ipython-input-11-243e9e4b4a0c> in <module>()
----> 1 df = DataFrame(index=mi, columns=mi)

../pandas/pandas/core/frame.py in __init__(self, data, index, columns, dtype, copy)
    395             mgr = self._init_mgr(data, index, columns, dtype=dtype, copy=copy)
    396         elif isinstance(data, dict):
--> 397             mgr = self._init_dict(data, index, columns, dtype=dtype)
    398         elif isinstance(data, ma.MaskedArray):
    399             mask = ma.getmaskarray(data)

../pandas/pandas/core/frame.py in _init_dict(self, data, index, columns, dtype)
    526
    527         return _arrays_to_mgr(arrays, data_names, index, columns,
--> 528                               dtype=dtype)
    529
    530     def _init_ndarray(self, values, index, columns, dtype=None,

../python/pandas/pandas/core/frame.py in _arrays_to_mgr(arrays, arr_names, index, columns, dtype)
   5703     axes = [_ensure_index(columns), _ensure_index(index)]
   5704
-> 5705     return create_block_manager_from_arrays(arrays, arr_names, axes)
   5706
   5707 def extract_index(data):

../pandas/pandas/core/internals.pyc in create_block_manager_from_arrays(arrays, names, axes)
   2232 def create_block_manager_from_arrays(arrays, names, axes):
   2233     try:
-> 2234         blocks = form_blocks(arrays, names, axes)
   2235         mgr = BlockManager(blocks, axes)
   2236         mgr._consolidate_inplace()

../pandas/pandas/core/internals.pyc in form_blocks(arrays, names, axes)
   2315
   2316     if len(object_items) > 0:
-> 2317         object_blocks = _simple_blockify(object_items, items, np.object_, is_unique=is_unique)
   2318         blocks.extend(object_blocks)
   2319

../pandas/pandas/core/internals.pyc in _simple_blockify(tuples, ref_items, dtype, is_unique)
   2334 def _simple_blockify(tuples, ref_items, dtype, is_unique=True):
   2335     """ return a single array of a block that has a single dtype; if dtype is not None, coerce to this dtype """
-> 2336     block_items, values, placement = _stack_arrays(tuples, ref_items, dtype)
   2337
   2338     # CHECK DTYPE?

../pandas/pandas/core/internals.pyc in _stack_arrays(tuples, ref_items, dtype)
   2395         items = _ensure_index([ n for n in names if n in ref_items ])
   2396         if len(items) != len(stacked):
-> 2397             raise Exception("invalid names passed _stack_arrays")
   2398
   2399     return items, stacked, placement

Exception: invalid names passed _stack_arrays

[jreback]

you don't like the error msg? (I am not sure i even remember what it's for!)

[jtratner]

@jreback :P

On Fri, Jun 28, 2013 at 7:07 PM, jreback notifications@github.com wrote:

you don't like the error msg? (I am not sure i even remember what it's
for!)

—
Reply to this email directly or view it on GitHubhttps://github.com//issues/4078#issuecomment-20219320
.

[cpcloud]

just got to the bottom of this: it only happens for a duplicate index value. works fine when there are no duplicates

[jreback]

does this fail for you on curent master?

[cpcloud]

yes but only when things aren't sorted and there are duplicates:

In [12]: tuples = [(3, 3), (2, 3), (3, 3)]

In [13]: mi = MultiIndex.from_tuples(tuples)

In [14]: df = DataFrame(index=mi,columns=mi)
Exception: invalid names passed _stack_arrays

while

In [15]: tuples = [(2, 3), (3, 3), (3, 3)]

In [16]: mi = MultiIndex.from_tuples(tuples)

In [17]: df = DataFrame(index=mi,columns=mi)

works fine

[jreback]

ahh...this is the case @jtratner talking about the other day...

[jreback]

ok....i'll just put tests up (and nto merge)

[cpcloud]

yes ... this happened during my refactor of read_html bc skiprows=1, header=[0, 1] and tupleize_cols=False are the trifecta of doom