排序熊猫数据框矢量化方式

Question

The sample input dataframe is as follows

df_input = pd.DataFrame([[1,2,3,4,5], [2,1,4,7,6], [5,6,3,7,np.nan], [np.nan,np.nan,np.nan,np.nan]], columns=["A", "B","C","D","E"])

Expected Output

df_output=pd.DataFrame([[-1,-1,0,1,1],[-1,-1,0,1,1],[-1,1,-1,1,0],[0,0,0,0,0]],columns=["A", "B","C","D","E"])

Here is what I am trying.
1) Sort every row.

2)Assign -1 to half of smallest valid observations and +1 to largest valid observations.

3) NaN needs to have zero.

4) In case of odd number of columns, the median needs to have zero.

The following code works well for even number of columns.

df_input[:] = np.where(df_input.rank(axis=1) > df_input.shape[1] / 2, 1, -1)

How do I expand this to odd number of columns and account for NaN. Thanks in advance.

Answer 1

I believe you need numpy.select :

a = df_input.rank(axis=1)
x = df_input.shape[1] / 2

m1 = a < x
m2 = a > x
m3 = (a.eq(a.mean(axis=1), axis=0))

df = pd.DataFrame(np.select([m3, m2, m1], [0, 1, -1], 0), columns=df_input.columns)
print (df)
   A  B  C  D  E
0 -1 -1  0  1  1
1 -1 -1  0  1  1
2 -1  1 -1  1  0
3  0  0  0  0  0

Answer 2

It seems that you are using a string for the NAN ('NAN').

then:

df_input = pd.DataFrame([[1,2,3,4,5], [2,1,4,7,6], [5,6,3,7,np.nan], [np.nan,np.nan,np.nan,np.nan]], columns=["A", "B","C","D","E"])

    df_input
        A   B   C   D   E
0   1.0 2.0 3.0 4.0 5.0
1   2.0 1.0 4.0 7.0 6.0
2   5.0 6.0 3.0 7.0 NaN
3   NaN NaN NaN NaN NaN

df2 = df_input.copy()
df2[:] = np.where(df2.isna(), 0, np.where(df2.rank(axis=1) > df2.shape[1] / 2, 1, -1))

df2

        A   B   C   D   E
0   -1  -1  1   1   1
1   -1  -1  1   1   1
2   -1  1   -1  1   0
3   0   0   0   0   0