Does np.array's astype prevent future edits in DataFrames? - numpy

Question

I can change the first entry of the DataFrame initially:
In [6]: df = pd.DataFrame(np.random.rand(5,2))
In [7]: df
Out[7]:
0 1
0 0.514592 0.459589
1 0.329704 0.409099
2 0.061246 0.966191
3 0.336747 0.908513
4 0.169220 0.468437
In [8]: df.ix[0][0] = 1
In [9]: df
Out[9]:
0 1
0 1.000000 0.459589
1 0.329704 0.409099
2 0.061246 0.966191
3 0.336747 0.908513
4 0.169220 0.468437
But after I do this:
In [10]: df[0] = np.floor(df.index / 10).astype(int) * 10
In [11]: df
Out[11]:
0 1
0 0 0.459589
1 0 0.409099
2 0 0.966191
3 0 0.908513
4 0 0.468437
I can't find a way to change it.
In [12]: df.ix[0][0] = 1
In [13]: df
Out[13]:
0 1
0 0 0.459589
1 0 0.409099
2 0 0.966191
3 0 0.908513
4 0 0.468437
And I can't even change elements from other columns
In [16]: df.ix[0][1] = 1
In [17]: df
Out[17]:
0 1
0 0 0.459589
1 0 0.409099
2 0 0.966191
3 0 0.908513
4 0 0.468437
What's up with this?

Answer 1

you are editing a copy, try
In [3]: df = pd.DataFrame(np.random.rand(5,2))
In [4]: df[0] = np.floor(df.index / 10).astype(int) * 10
In [5]: df
Out[5]:
0 1
0 0 0.201611
1 0 0.390364
2 0 0.727422
3 0 0.941035
4 0 0.036764
In [6]: df.ix[0,1] = 1
In [7]: df
Out[7]:
0 1
0 0 1.000000
1 0 0.390364
2 0 0.727422
3 0 0.941035
4 0 0.036764