错误：不再支持将类似列表的标签传递给.loc或[]且缺少任何标签

Question

发现此代码段基本上可以增加我的负面评论，从而更好地训练我的模型。当我通过运行它时，出现此错误。看起来在idx周围。有人对此有好的解决方案吗？

不再支持将缺少列表的标签传递给.loc或[]，并且缺少任何标签 https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#deprecate-loc-reindex-listlike'

from sklearn.utils import shuffle
import numpy as np

labels, num = np.unique(y_train, return_counts=True)
#print(labels)
u=min(labels)
intial = 1

#set the desired size of the oversampled cells
maxcnt = np.int(max(num)/2)

for labl, n in zip(labels, num):
    x0 = X_train[y_train==labl]
    y0 = y_train[y_train==labl]
    # print (x0)

    remain = maxcnt
    print (remain)
    while remain >= n;
        if label == u and initial == 1;
            X_Train = x0
            y_Train = y0
            remain -= n
            initial = 0
        else:
            X_Train = np.concatenate((X_Train, x0), axis=0)
            y_Train = np.concatenate((y_Train, y0), axis=0)
            remain -= n
    if remain > 0 and remain < n:
        idx = np.random.choice(np.arange(len(y0)), remain, replace=False)
       #print(idx)
        X_Train = np.concatenate((X_Train, x0[idx]), axis=0)
        y_Train = np.concatenate((y_Train, y0[idx]), axis=0)
        remain -= n
X_Train, y_Train = shuffle(X_Train, y_Train)
np.unique(X_Train, return_counts=True)