2020-03-31

To be continued(自用）

今天在处理数据的时候，得到一个包含缺失值个数的series，想把其中缺失个数大于某个数的索引提取出来

其实完全不用这么写

说明不能只是[df_1.isnull().sum().sort_values(ascending=False).values]>70
而是要构造一个形状一致全为70的列表

na_cate = df_1.isnull().sum().sort_values(ascending=False)
na_cate

F2-GrayLevelCooccurenceMatrix39-7Correlation                84
F2-GrayLevelCooccurenceMatrix39-7InformationMeasureCorr1    84
F2-GrayLevelCooccurenceMatrix39-7Entropy                    83
F2-GrayLevelCooccurenceMatrix39-7InformationMeasureCorr2    83
F2-GrayLevelCooccurenceMatrix39-7SumVariance                83
                                                            ..
F2-GrayLevelCooccurenceMatrix37-1InverseDiffMomentNorm       0
F2-GrayLevelCooccurenceMatrix37-4InverseDiffMomentNorm       0
F2-GrayLevelCooccurenceMatrix37-7InverseDiffMomentNorm       0
F2-GrayLevelCooccurenceMatrix38-1InverseDiffMomentNorm       0
ID                                                           0
Length: 1426, dtype: int64

delete_columns_list = na_cate[na_cate.values>70].index.tolist()
len(delete_columns_list)

63

df_1.drop(delete_columns_list,axis=1)