Initializing DataFrom from list of OrderedDicts does not preserve Column order

[cod3monk]

>>> import pandas as pd
>>> from collections import OrderedDict
>>> od = [OrderedDict([('z', 0), ('a', 1), ('*', 2)])]
>>> pd.DataFrame(od)
   *  a  z
0  2  1  0
>>> pd.DataFrame.from_dict(od)
   *  a  z
0  2  1  0

If OrderedDicts are used, I would expect the column order of the first OrderDict that is encountered to be preserved, as the following:

>>> pd.DataFrame(od, columns=od[0].keys())
   z  a  *
0  0  1  2
>>>

If different orders occur in the list, the earlier would overwrite the later. If new columns are introduced with subsequent list entries, I would expect all new columns to be appended to the original columns, with their relative order preserved.

[jreback]

this is a duplicate of ##13304

which is already in master and will be in 0.19.0, see here

[cod3monk]

Thanks and sorry for the duplicate.