在 Pandas 中尋找最接近的值

Question

X     A    B   C    D 
10  10.2  16   28  30 
20  27    15   14  16
30  30.4  34   35  45

我想創建一個新列，顯示來自 A、B、C 或 D 的值，以防它們接近“X”的 2%。

預期結果：

X     A    B   C    D   result
10  10.2  16   28  30   10.2
20  27    15   14  16   NaN
30  30.4  34   35  45   30.4

我可以使用以下代碼識別接近“X”的值，但現在確定如何從 A、B、C、D 列中取出值：

cond = np.isclose(df.X, df['A'], rtol=0.02) | np.isclose(df.X, df['B'], rtol=0.02) | np.isclose(df.X, df['C'], rtol=0.02) | np.isclose(df.X, df['D'], rtol=0.02)

df['result'] = np.where(cond,#See note,np.nan)

#note：如何放置符合條件的列值？

很少有幫助將不勝感激。 謝謝！

編輯：

ID  OP     X   Vl  Ch  A    B   C    D   result
          10           10.2  16   28  30   10.2
          20           27    15   14  16   NaN
          30           30.4  34   35  45   30.4

空白值中有一些數據。 ID 是字符串，其他都是浮點數。

Answer 1

In [1]: df = pd.DataFrame([[10, 10.2, 16, 28, 30], [20, 27, 15, 14, 16], [30, 30.4, 34, 25, 35]], columns=["X", "A", "B", "C", "D"])

In [2]: def f(x):
     ...:     # If there are multiple values that meet the conditions, please modify here
     ...:     c = None  # []
     ...:     for i in x[1:]:
     ...:         if i/x[0] >=1.0 and i/x[0] <=1.02:
     ...:             c = i
     ...:             break
     ...:             # c.append(i)
     ...:     return c
     ...:

In [3]: df['result'] = df.apply(lambda x: f(x), axis=1)

In [4]: df
Out[5]:
    X     A   B   C   D  result
0  10  10.2  16  28  30    10.2
1  20  27.0  15  14  16     NaN
2  30  30.4  34  25  35    30.4

Answer 2

val = [[10,  10.2,  16,   28,  30], [20, 27,  15,  14, 16], [30, 30.4,  34,  35, 45]]
df = pd.DataFrame(val, columns=list('XABCD'))

# split df for convenience:
target_X = df['X']
df2 = df.drop('X', axis=1)
tol = 0.02

# get the closest value to X within the tolerance tol:
df['Result'] = df2[df2.subtract(target_X, axis=0).abs().apply(lambda x: x <= target_X * tol)].min(1)
print(df)