客趣旅游网
1.数据清洗的代码:
import pandas as pd import numpy as np # 创建空的df,保存测试数据 test_df = pd.DataFrame({'K1':['C1','C1','C2','C3','C4','C2','C1'],'K2':['A','A','B','C','D',np.NaN,np.NaN]}) # 按K1列进行分组,组内进行unique操作(去除重复元素,返回元组或列表) test_df_unique = pd.DataFrame(test_df.groupby(['K1'])['K2'].agg('unique')) # 自定义函数判断元组中是否含有nan def has_nan(list): flag = False for x in list: if x is np.NaN: flag = True break return flag # 自定义函数判断元组中是否不含有nan def no_nan(list): flag = True for x in list: if x is np.NaN: flag = False break return flag # 获取k2列含有nan的数据 test_df_unique_has_nan = test_df_unique[test_df_unique['K2'].apply(has_nan)] # 获取k2列不含有nan的数据 test_df_unique_no_nan = test_df_unique[test_df_unique['K2'].apply(no_nan)] # 管理测试数据,获取源数据 test_df_get = test_df[test_df['K1'].isin(test_df_unique_has_nan.index.tolist())] test_df_alone = test_df[test_df['K1'].isin(test_df_unique_no_nan.index.tolist())] # 去除含nan的重复数据 test_df_get_nonan = test_df_get[~test_df_get['K2'].isna()] # 组合数据 result = test_df_get_nonan.append(test_df_alone) # 去重,得到最终结果 result_save = result.drop_duplicates(subset=['K1','K2'],keep='last') # 结果落地 result_save.to_excel('C:/Users/zhen/Desktop/数据清洗之去重.xlsx')
2、测试数据:
3、结果: