熊猫数据框字典：如何知道是引用字典中的数据框副本还是引用字典中的数据框？

Question

Hi I've run into some interesting behavior that I can't find an explanation.

I have a dict called combinedDict. It has string keys and its elements are pandas dataframes.

I want to select the dataframe named 'early'. I create a variable a equal to that dataframe. I then want to edit the ID column in that dataframe by appending the string '_early' to each row of that column. I do this using the following code:

a = combinedDict['early']
a['ID'] = [(s + '_early') for s in a['ID'].tolist()]

When I do this, the string '_early' is appended to every row of the column in dataframe a but it is also appended to every row of the dataframe stored in combinedDict['early'].

On the other hand, when I use the following code, adding .copy(), only dataframe a is affected while the one stored in combinedDict['early'] is not. This is the first time I've run into this behavior. Is this just a feature of pandas dataframes?

Answer 1

Let's try this:

In [87]: df1 = pd.DataFrame({'a': [1,2,3,4,5,6,7,8,9]})

In [88]: df1
Out[88]:
   a
0  1
1  2
2  3
3  4
4  5
5  6
6  7
7  8
8  9

In [89]: df2 = df1

In [90]: id(df1) == id(df2)
Out[90]: True

In [91]: df2.ix[df2.a > 4, 'a'] = 0

In [92]: df1
Out[92]:
   a
0  1
1  2
2  3
3  4
4  0
5  0
6  0
7  0
8  0

So your a is a reference to combinedDict['early']

Here is an excerpt from the documentation :

Mutability and copying of data

All pandas data structures are value-mutable (the values they contain can be altered) but not always size-mutable. The length of a Series cannot be changed, but, for example, columns can be inserted into a DataFrame. However, the vast majority of methods produce new objects and leave the input data untouched. In general, though, we like to favor immutability where sensible.