import logging import pandas as pd from sqlalchemy import create_engine # 创建数据库引擎 engine235 = create_engine("postgresql+psycopg2://postgres:Goojia123456@192.168.0.235:5432/fangjia?client_encoding=utf8", echo=False) # 执行SQL查询获取区域价格数据 region_sql = "SELECT DISTINCT urbn_nm,gxda_admndiv_nm FROM price_stat.avm_district_price_copy1" region_data = pd.read_sql(region_sql, engine235) # 假设这是你的pgsql数据 all_region_data = pd.DataFrame() pd.set_option('mode.chained_assignment', None) def region_increase(region_data, all_region_data): # 遍历区域列表,对每个区域进行查询 for index, row in region_data.iterrows(): city = row['urbn_nm'] region = row['gxda_admndiv_nm'] query = f"SELECT urbn_nm as 城市, gxda_admndiv_nm as 区域, gujiabds_clc_yrmo_val as 月份, gujia_bds_clc_avprc as 估家计算均价, gujia_bds_urbn_avprc as 估家城市均价 FROM price_stat.avm_district_price_copy1 WHERE urbn_nm = '{city}' AND gxda_admndiv_nm = '{region}'" region_data = pd.read_sql(query, engine235) if region_data.empty: continue # 复制一份原始的日期列,用于计算涨跌幅时保持原始数据不变 region_data['original_date'] = region_data['月份'] # 将日期列转换为日期类型 region_data['月份'] = pd.to_datetime(region_data['月份'], format='%Y%m') # 对价格数据按照月份排序 region_data.sort_values(by='月份', inplace=True) # 计算涨跌幅 region_data['涨跌幅'] = region_data['估家计算均价'].diff() / region_data['估家计算均价'].shift(1) * 100 # 计算涨跌幅,并保留两位小数,展示为百分比形式 price_change = region_data['估家计算均价'].pct_change() * 100 region_data['涨跌幅'] = (price_change[price_change != 0].fillna(0).round(2).astype(str) + '%').replace('0.0%', '') # 将原始日期列替换为计算涨跌幅后的日期列 region_data['月份'] = region_data['original_date'] # 删除临时列 region_data.drop(columns=['original_date'], inplace=True) # 将区域数据添加到总的数据中 all_region_data = pd.concat([all_region_data, region_data], ignore_index=True) return all_region_data def region_missing(): df = region_increase(region_data, all_region_data) df['月份1'] = df['月份'] # 将日期字符串转换为日期时间对象 df["月份"] = pd.to_datetime(df["月份"], format="%Y%m") # 获取当前时间 current_time = pd.Timestamp.today().replace(hour=0, minute=0, second=0, microsecond=0) # 计算前一年到上月的时间范围 one_year_ago = current_time - pd.DateOffset(years=1, days=current_time.day - 1) last_month = current_time - pd.DateOffset(months=1, days=current_time.day - 1) # 创建一个包含所有可能月份的日期范围 all_months = pd.date_range(start=one_year_ago, end=last_month, freq="MS") # 获取每个城市-区域组合缺失的月份 missing_months = [] for (city, region), group in df.groupby(["城市", "区域"]): available_months = group["月份"] missing = all_months.difference(available_months) if len(missing) > 0: missing_str = ",".join([m.strftime("%Y%m") for m in missing]) missing_months.append({"城市": city, "区域": region, "缺失月份": missing_str}) if missing_months: # 创建DataFrame来存储结果 missing_df = pd.DataFrame(missing_months) path = 'E:\\goojia\\file\\excel文件\\airflow\\missing_region.xlsx' missing_df.to_excel(path, index=False) # 如果区域在 missing_df 中,则将 df 中的该区域数据删除 for city, region in zip(missing_df['城市'], missing_df['区域']): df = df[(df['城市'] != city) | (df['区域'] != region)] df['月份'] = df['月份1'] # 删除月份1列 df.drop(columns=['月份1'], inplace=True) output_path = 'E:\\goojia\\file\\excel文件\\airflow\\all_region_price_with_change.xlsx' df.to_excel(output_path, index=False) logging.info(f"区域涨跌幅数据已写入Excel文件:{output_path}") region_missing()