ENH/API: allow DataFrame constructor to better accept list-like collections (GH3783,GH4297) #4829
Conversation
| @@ -418,7 +418,8 @@ def __init__(self, data=None, index=None, columns=None, dtype=None, | |||
| if index is None and isinstance(data[0], Series): | |||
| index = _get_names_from_index(data) | |||
|
|
|||
| if isinstance(data[0], (list, tuple, collections.Mapping, Series)): | |||
| if isinstance(data[0], (list, tuple, collections.Mapping, | |||
| collections.Sequence, Series)) and is_list_like(data[0]): | |||
There was a problem hiding this comment.
why couldn't you just remove the entire type check and just use is_list_like instead here?
There was a problem hiding this comment.
DataFrame([DataFrame(dict(A = lrange(5)))])
this beauty here (which if you use just is_list_like) collapses it...maybe should just completely disallow this
There was a problem hiding this comment.
I think core.common.is_list_like(DataFrame(dict(A = lrange(5)))) returns True.
There was a problem hiding this comment.
that's the problem, it needs to be initialized to an object array, rather than like a list-of-lists which is intialized differently
There was a problem hiding this comment.
ah - I see...could you check for PandasObjects first or something? (or add
not isinstance(data[0], (DataFrame, Panel))?) otherwise I guess it's fine
On Thu, Sep 12, 2013 at 9:08 PM, jreback [email protected] wrote:
In pandas/core/frame.py:
@@ -418,7 +418,8 @@ def init(self, data=None, index=None, columns=None, dtype=None,
if index is None and isinstance(data[0], Series):
index = _get_names_from_index(data)
if isinstance(data[0], (list, tuple, collections.Mapping, Series)):if isinstance(data[0], (list, tuple, collections.Mapping,collections.Sequence, Series)) and is_list_like(data[0]):that's the problem, it needs to be initialized to an object array, rather
than like a list-of-lists which is intialized differently—
Reply to this email directly or view it on GitHubhttps://github.com//pull/4829/files#r6339304
.
There was a problem hiding this comment.
no good solution either way; its basically checking if the input is 1-d or 2-d, hmm....let me try something..
There was a problem hiding this comment.
eh, I'm just not a fan of collections in general, because it's slowish, but this is probably the better option because it's more explicit, and, as @y-p notes, allows for objects to register themselves.
|
@jtratner that worked....now not dependent on the type per se only that its list-like and 1dim |
…g. list of
`ollections.Sequence and array.Array objects (GH3783 and GH42971)
… construction ENH: add generator support in DataFrame constructor
|
👍 |
ENH/API: allow DataFrame constructor to better accept list-like collections (GH3783,GH4297)
closes #3783
closes #4297
related to #2305 (as now accept GeneratorType), but just convert to list, don't incrementally create the frame