Better error messages

[jankatins]

I'm trying to build a dataframe from a list of list and I get an "AssertionError" and I have no idea what I did wrong or what I should do differently. A shorted version of the data and code:

pprint(_fields)
[ 'name',
'timescited',
'weight',
'closeness',
'betweenness',
'degree',
'numberofworks',
'pagerank',
'pages',
'constraint',
'citationaverage',
'eigenvector']

pprint(_data2)
[ ['Huselid, Ma', 'Gulati, R', 'Damanpour, F', 'Mcallister, Dj', 'Tsai, Wp'],
[2721, 5251, 1269, 1287, 2834],
[6, 17, 6, 3, 6],
[ 0.0002510562854116362,
0.00025108339541739277,
0.00025105435150127433,
0.00025106898828104144,
0.0002510723039607744],
[ 23311.0,
173596.65383408728,
79279.82282582425,
18701.108026093425,
57261.74716265692],
[4, 11, 9, 3, 7],
[6, 17, 6, 3, 6],
[ 0.0001438440250079284,
0.0003063098736173118,
0.00027071793856870986,
7.839463995668608e-05,
0.00020047003509480145],
[130, 387, 148, 68, 88],
[ 0.28573421556122447,
0.12411566040831735,
0.19959178902739988,
0.3703749360800643,
0.200788126303937],
[453, 308, 211, 429, 472],
[ 5.889013732612275e-08,
0.0005066654664776551,
4.567816673280219e-07,
2.4226314239685797e-06,
4.069482619394397e-07]]

df = DataFrame(_data2, columns=_fields )
Traceback (most recent call last):
File "", line 1, in
File "C:\portabel\Python27\lib\site-packages\pandas\core\frame.py", line 208, in init
copy=copy)
File "C:\portabel\Python27\lib\site-packages\pandas\core\frame.py", line 273, in _init_ndarray
block = make_block(values.T, columns, columns)
File "C:\portabel\Python27\lib\site-packages\pandas\core\internals.py", line 211, in make_block
do_integrity_check=do_integrity_check)
File "C:\portabel\Python27\lib\site-packages\pandas\core\internals.py", line 26, in init
assert(len(items) == len(values))
AssertionError

It would be nice if pandas could output a more descriptive error message in case some inputs are not what pandas expects.

Thanks a lot for the lib!

Jan

[jankatins]

I found the problem after some try&error: _data2 needs to be a dict... The problem about the unhelpfull error message remains :-)

[wesm]

ok, here's the new exception:

In [3]: DataFrame(data2, columns=fields)
---------------------------------------------------------------------------
AssertionError                            Traceback (most recent call last)
/home/wesm/code/pandas/<ipython-input-3-7de2f1fbdd32> in <module>()
----> 1 DataFrame(data2, columns=fields)

/home/wesm/code/pandas/pandas/core/frame.pyc in __init__(self, data, index, columns, dtype, copy)
    209         elif isinstance(data, list):
    210             if len(data) > 0 and isinstance(data[0], (list, tuple)):
--> 211                 data, columns = _list_to_sdict(data, columns)
    212                 mgr = self._init_dict(data, index, columns, dtype=dtype)
    213             else:

/home/wesm/code/pandas/pandas/core/frame.pyc in _list_to_sdict(data, columns)
   3557         if len(columns) != len(content):
   3558             raise AssertionError('%d columns passed, passed data had %s '
-> 3559                                  'columns' % (len(columns), len(content)))
   3560 
   3561     sdict = dict((c, lib.maybe_convert_objects(vals))

AssertionError: 12 columns passed, passed data had 5 columns

for this data you'd want to do:

In [7]: DataFrame(zip(*data2), columns=fields)
Out[7]: 
<class 'pandas.core.frame.DataFrame'>
Int64Index: 5 entries, 0 to 4
Data columns:
name               5  non-null values
timescited         5  non-null values
weight             5  non-null values
closeness          5  non-null values
betweenness        5  non-null values
degree             5  non-null values
numberofworks      5  non-null values
pagerank           5  non-null values
pages              5  non-null values
constraint         5  non-null values
citationaverage    5  non-null values
eigenvector        5  non-null values
dtypes: int64(6), float64(5), object(1)

because you have the columns in the rows, zip(*data) transposes a list of lists

[jankatins]

Thanks a lot for this change!

I'm still not sure if I would have understood the error. How about something like "12 column name(s) passed, but passed data had 5 columns." or even a added"... If you passed in a list of list, maybe transposing the data with zip(*data) helps."?