从pandas DataFrame中选择行的问题

Question

I encounter an error when I tried to extract rows from DataFrame where the rows match the conditions.
The code I used is a very simple one:

    for c in classes:
           print(X[y == c])

where X is features of samples in DataFrame, y is classes of the samples in DataFrame, and c is a list of class labels.
I used this code for two data sets, and it works with a dataset, but does not with the other, although the both are formatted in the same way.

With the data set I had an error, I printed y == c by:

     print(y == c)

and it returned this:

           Classes
     0     True
     1     True
     2     True
           ...
     4572  False
     4573  False
     4574  False

Therefore, I am assuming that the condition matching is working properly.
However, when I print X[y == c] by:

  print(X[y == c])

the result is like this:

            0   1   2
     0    NaN NaN NaN
     1    NaN NaN NaN
     2    NaN NaN NaN
     3    NaN NaN NaN
           ...
     4574 NaN NaN NaN

To note, the X and y are ordinary DataFrame that looks like these:
X

                     0           1          2
     0       -3.786900    9.411757  -2.246594
     1      742.632101  -74.001353  -0.567936
     2     2019.854074  102.077111 -23.776775
     3      -93.048341    3.008569  -1.043599
           ...
     4754  -247.754953   -6.851270  -0.984777

y

            Classes
     0      0
     1      0
     2      0
            ...
     4572   2
     4573   2
     4574   2

Can this problem be sorted out?

Answer 1

如果在数据和相同的索引值的相同长度DataFrames使用DataFrame.eq通过柱Classes沿0轴布尔DataFrame ，然后检查所述至少一个True每行由DataFrame.any和由滤波器boolean indexing ：

df = X[X.eq(Y['Classes'], axis=0).any(axis=1)]

Answer 2

I figured out the cause of the problem.
X and y are the same length. I had the above problem only with Data1, but not with Data2.

I checked the data type of y in Data1 and Data2, and found that y was

• < class 'pandas.core. series.Series > with Data2 (working)
• < class 'pandas.core. frame.DataFrame '> with Data1 (not working)

I converted y in Data1 into Series by:

    y = y.ix[:,0]

,then extraction of rows came to work properly.