Reputation: 525
I am writing a program in Python that opens input files, does some simple text parsing, and outputs. The input comes as an ASCII file with several similarly formatted blocks of text. So, I thought I'd use this as an opportunity to get the hang of defining my own classes.
I have a parent class, pFrame
, that I want to inherit the properties of the pandas.DataFrame
class. Since my input text files contain two similar (but not identical) types of columnar text, I define two other classes (pFrameA
and pFrameB
) that each inherit the parent class. At present the child classes just initialize a few variables; later I can define simple helper methods for one, the other, or both classes as needed.
Here's a trimmed down version of the module I wrote to define these classes:
import pandas as pd
class pFrame(pd.DataFrame):
pass
class pFrameA(pFrame):
def __init__(self):
self.units = ('msec', 'psi')
self.numFormat = (int, float)
self._gcHeaderStr = " Time Cell"
class pFrameB(pFrame):
def __init__(self):
self.units = ('in', 'in')
self.numFormat = (float, float)
self._gcHeaderStr = "Disp 1 Disp 2"
But when I try to test out these class definitions, Python goes into an endless recursion loop:
>>> import pFrameModule
>>> p=pFrameModule.pFrameA()
...
...
if name in self.columns:
File "properties.pyx", line 55, in pandas.lib.AxisProperty.__get__ (pandas/lib.c:29248)
File "/usr/lib/python2.7/dist-packages/pandas/core/frame.py", line 2085, in __getattr__
if name in self.columns:
File "properties.pyx", line 55, in pandas.lib.AxisProperty.__get__ (pandas/lib.c:29248)
File "/usr/lib/python2.7/dist-packages/pandas/core/frame.py", line 2085, in __getattr__
if name in self.columns:
File "properties.pyx", line 55, in pandas.lib.AxisProperty.__get__ (pandas/lib.c:29248)
RuntimeError: maximum recursion depth exceeded
Can someone give me a quick pointer to understand what I'm missing?
Upvotes: 1
Views: 176
Reputation: 64258
Modifying your code to the following appears to fix it:
import pandas as pd
class pFrame(pd.DataFrame):
def __init__(self):
super(pFrame, self).__init__()
class pFrameA(pFrame):
def __init__(self):
super(pFrameA, self).__init__()
self.units = ('msec', 'psi')
self.numFormat = (int, float)
self._gcHeaderStr = " Time Cell"
class pFrameB(pFrame):
def __init__(self):
super(pFrameB, self).__init__()
self.units = ('in', 'in')
self.numFormat = (float, float)
self._gcHeaderStr = "Disp 1 Disp 2"
p = pFrameA()
print p._gcHeaderStr # To prove something is happening
I suspect that by neglecting the call to super
in each of your constructors, you weren't properly initializing the DataFrame class, causing it to break under the scenes.
Upvotes: 2