import pandas as pd # 计算每列的完整率 completeness = (1 - df.isnull().sum() / len(df)) * 100 print('各列完整率(%):') print(completeness) # 计算整表的完整率 total_cells = df.shape[0] * df.shape[1] missing_cells = df.isnull().sum().sum() overall = (1 - missing_cells / total_cells) * 100 print(f'整体完整率: {overall:.2f}%') # 完整率低于90%的列标红 low_cols = completeness[completeness < 90].index.tolist() print(f'完整率低于90%的列: {low_cols}')

# 规则1：手机号必须是11位数字 phone_valid = df['手机号'].astype(str).str.match(r'^\d{11}$') print(f'手机号有效比例: {phone_valid.mean()*100:.2f}%') # 规则2：邮箱必须包含@ email_valid = df['邮箱'].str.contains('@', na=False) print(f'邮箱有效比例: {email_valid.mean()*100:.2f}%') # 规则3：年龄必须在合理范围 age_valid = (df['年龄'] >= 0) & (df['年龄'] <= 120) print(f'年龄合理比例: {age_valid.mean()*100:.2f}%') # 综合准确率 accuracy = (phone_valid & email_valid & age_valid).mean() * 100 print(f'综合准确率: {accuracy:.2f}%')

# 检查日期一致性：注册日期不应晚于最后登录日期 inconsistent = df[df['注册日期'] > df['最后登录日期']] print(f'日期不一致记录数: {len(inconsistent)}') # 检查枚举值一致性：性别只能是男/女 valid_gender = df['性别'].isin(['男', '女']) print(f'性别规范比例: {valid_gender.mean()*100:.2f}%') # 检查金额一致性：订单金额 = 单价 * 数量 df['计算金额'] = df['单价'] * df['数量'] amount_match = (df['订单金额'] - df['计算金额']).abs() < 0.01 print(f'金额一致比例: {amount_match.mean()*100:.2f}%')

from datetime import datetime # 计算数据年龄（距离现在多久） now = datetime.now() df['数据年龄_天'] = (now - pd.to_datetime(df['采集时间'])).dt.days # 统计各年龄段的数据量 print(df['数据年龄_天'].value_counts().sort_index()) # 及时性指标：24小时内的数据占比 timely = (df['数据年龄_天'] <= 1).mean() * 100 print(f'24小时内数据占比: {timely:.2f}%')

report = { '总记录数': len(df), '总字段数': len(df.columns), '整体完整率(%)': round((1 - df.isnull().sum().sum() / (df.shape[0]*df.shape[1])) * 100, 2), '重复记录数': df.duplicated().sum(), '重复率(%)': round(df.duplicated().mean() * 100, 2), } for k, v in report.items(): print(f'{k}: {v}') # 保存报告 pd.DataFrame([report]).to_csv('quality_report.csv', index=False)