python测试工具: 实现数据源自动核对

测试业务需要：

现有A系统作为下游数据系统，上游系统有A1,A2,A3...

需要将A1,A2,A3...的数据达到某条件后（比如：A1系统销售单提交出库成功）自动触发MQ然后再经过数据清洗落到A系统，并将清洗后数据通过特定规则汇总在A系统报表中

现在需要QA同学验证的功能是：

A系统存储数据清洗后的库表(为切片表)有几十个，且前置系统较多，测试工作量也较多

需要核对清洗后A存库数据是否准确

清洗规则：（1）直接取数 （2）拼接取数 （3）映射取数

直接取数字段在2系统表中字段命名规则一致

so,以下测试工具是针对直接取数规则来开发，以便于测试

代码实现步骤：

（1）将表字段，来源系统表和切片表 数据库链接信息，查询字段 作为变量

将这些信息填入input.xlsx 作为入参

（2）读取表字段，根据来源系统表 数据库链接信息，查询字段

查询来源库表，将查询出字段值存储outfbi.xlsx

  (3)读取表字段，根据切片表 数据库链接信息，查询字段

查询切片库表，将查询出字段值存储outods.xlsx

（4）对比outfbi.xlsx，outods.xlsx的字段值

对比后生成result.xlsx文件，新增列校验结果

核对字段值一致校验结果为Success,否则为Fail

代码如下：

入参文件见附件

DbcheckApi.py

import os
import pymysql
import pandas as pd
from openpyxl import load_workbook
from openpyxl.styles import PatternFill
import datetime
import ast"""测试数据路径管理"""
SCRIPTS_DIR = os.path.dirname(os.path.dirname(os.path.abspath(__file__)))
GENERATECASE_DIR = os.path.join(SCRIPTS_DIR, "dbcheck")
inputDATAS_DIR = os.path.join(GENERATECASE_DIR, "inputdata")
outDATAS_DIR = os.path.join(GENERATECASE_DIR, "outdata")class DbcheckApi():def __init__(self,data):self.inputexcel=dataworkbook = load_workbook(filename=self.inputexcel)sheet = workbook['数据源']# 读取来源表-连接信息sourcedb_connection_info = ast.literal_eval(sheet['B3'].value)odsdb_connection_info = ast.literal_eval(sheet['B4'].value)source_db = sheet['C3'].value.strip()ods_db = sheet['C4'].value.strip()source_queryby = sheet['D3'].value.strip()ods_queryby = sheet['D4'].value.strip()print(sourcedb_connection_info)print(odsdb_connection_info)print(source_db)print(ods_db)print(source_queryby)print(ods_queryby)self.sourcedb = sourcedb_connection_infoself.odsdb = odsdb_connection_infoself.source_db = source_dbself.ods_db = ods_dbself.source_queryby = source_querybyself.ods_queryby = ods_querybydef source_select_db(self):host = self.sourcedb.get('host')port = self.sourcedb.get('port')user = self.sourcedb.get('user')passwd = self.sourcedb.get('passwd')db = self.sourcedb.get('db')if not host or not port or not user or not passwd or not db:error_msg = "连接信息不完整"return {"code": -1, "msg": error_msg, "data": ""}cnnfbi = pymysql.connect(host=host,port=port,user=user,passwd=passwd,db=db)cursor = cnnfbi.cursor()try:# 读取Excel文件df = pd.read_excel(self.inputexcel, sheet_name='Sheet1')# 获取第1列，从第2行开始读取的字段名fields = df.iloc[1:, 0].tolist()print(fields)# 构建查询SQL语句sql = "SELECT {} FROM {} WHERE {}".format(', '.join(fields), self.source_db, self.source_queryby)print(sql)# 执行查询语句cursor.execute(sql)except pymysql.err.OperationalError as e:error_msg = str(e)if "Unknown column" in error_msg:column_name = error_msg.split("'")[1]msg={"code": -1, "msg": f"列字段 {column_name} 在 "+self.source_db+" 表结构中不存在，请检查！", "data": ""}print(msg)return {"code": -1, "msg": f"列字段 {column_name} 在 "+self.source_db+" 表结构中不存在，请检查！", "data": ""}else:return {"code": -1, "msg": error_msg, "data": ""}print(error_msg)# 获取查询结果result = cursor.fetchall()# 关闭游标和连接cursor.close()cnnfbi.close()# 检查查询结果是否为空if not result:return {"code": -1, "msg": f"查询无数据，请检查sql: {sql}", "data": ""}else:# 将结果转换为DataFrame对象df = pd.DataFrame(result, columns=fields)odskey=self.source_db+'表-字段'odsvalue=self.source_db+'表-字段值'# 创建新的DataFrame对象，将字段和对应值放在两列df_new = pd.DataFrame({odskey: fields, odsvalue: df.iloc[0].values})outexcel = os.path.join(outDATAS_DIR,  'outputfbi.xlsx')# 导出结果到Excel文件df_new.to_excel(outexcel, index=False)def ods_select_db(self):host = self.odsdb.get('host')port = self.odsdb.get('port')user = self.odsdb.get('user')passwd = self.odsdb.get('passwd')db = self.odsdb.get('db')if not host or not port or not user or not passwd or not db:raise ValueError("连接信息不完整")cnnfbi = pymysql.connect(host=host,port=port,user=user,passwd=passwd,db=db)cursor = cnnfbi.cursor()try:# 读取Excel文件df = pd.read_excel(self.inputexcel, sheet_name='Sheet1')# 获取第1列，从第2行开始读取的字段名fields = df.iloc[1:, 0].tolist()print(fields)# 构建查询SQL语句sql = "SELECT {} FROM {} WHERE {}".format(', '.join(fields), self.ods_db, self.ods_queryby)print(sql)# 执行查询语句cursor.execute(sql)except pymysql.err.OperationalError as e:error_msg = str(e)if "Unknown column" in error_msg:column_name = error_msg.split("'")[1]return {"code": -1, "msg": f"列 {column_name} 不存在"+self.ods_db+" 表结构中，请检查！", "data": ""}else:return {"code": -1, "msg": error_msg, "data": ""}# 获取查询结果result = cursor.fetchall()# 关闭游标和连接cursor.close()cnnfbi.close()# 将结果转换为DataFrame对象df = pd.DataFrame(result, columns=fields)# 创建新的DataFrame对象，将字段和对应值放在两列odskey=self.ods_db+'表-字段'odsvalue=self.ods_db+'表-字段值'df_new = pd.DataFrame({odskey: fields, odsvalue: df.iloc[0].values})# 导出结果到Excel文件outexcel = os.path.join(outDATAS_DIR,  'outputfms.xlsx')df_new.to_excel(outexcel, index=False)def check_order(self):self.source_select_db()self.ods_select_db()outputfbi = os.path.join(outDATAS_DIR,  'outputfbi.xlsx')outputfms = os.path.join(outDATAS_DIR,  'outputfms.xlsx')df_a = pd.read_excel(outputfbi)df_b = pd.read_excel(outputfms)# 创建新的DataFrame对象用于存储C表的数据df_c = pd.DataFrame()# 将A表的列写入C表for col in df_a.columns:df_c[col] = df_a[col]# 将B表的列���入C表for col in df_b.columns:df_c[col] = df_b[col]odsvalue=self.ods_db+'表-字段值'fbivalue=self.source_db+'表-字段值'# 比对A2和B2列的值，如果不一致，则在第5列写入"校验失败"df_c['校验结果'] = ''for i in range(len(df_c)):if pd.notnull(df_c.at[i, fbivalue]) and pd.notnull(df_c.at[i, odsvalue]):fbivalue_rounded = df_c.at[i, fbivalue]odsvalue_rounded = df_c.at[i, odsvalue]if isinstance(fbivalue_rounded, (int, float)):fbivalue_rounded = round(fbivalue_rounded, 3)elif isinstance(fbivalue_rounded, datetime.datetime):fbivalue_rounded = round(fbivalue_rounded.timestamp(), 3)else:try:fbivalue_rounded = round(float(fbivalue_rounded), 3)except ValueError:passif isinstance(odsvalue_rounded, (int, float)):odsvalue_rounded = round(odsvalue_rounded, 3)elif isinstance(odsvalue_rounded, datetime.datetime):odsvalue_rounded = round(odsvalue_rounded.timestamp(), 3)else:try:odsvalue_rounded = round(float(odsvalue_rounded), 3)except ValueError:passif fbivalue_rounded != odsvalue_rounded:df_c.at[i, '校验结果'] = 'Fail'else:df_c.at[i, '校验结果'] = 'Success'# 将结果写入到C.xlsx文件df_c.to_excel('checkhead_result.xlsx', index=False)# 打开C.xlsx文件并设置背景色book = load_workbook('checkhead_result.xlsx')writer = pd.ExcelWriter('checkhead_result.xlsx', engine='openpyxl')writer.book = book# 获取C.xlsx的工作表sheet_name = 'Sheet1'ws = writer.book[sheet_name]# 设置背景色为红色red_fill = PatternFill(start_color='FFFF0000', end_color='FFFF0000', fill_type='solid')# 遍历校验结果列，将不一致的单元格设置为红色背景for row in ws.iter_rows(min_row=2, min_col=len(df_c.columns), max_row=len(df_c), max_col=len(df_c.columns)):for cell in row:if cell.value == 'Fail':cell.fill = red_fill# 保存Excel文件writer.save()writer.close()if __name__ == '__main__':inputexcel = os.path.join(inputDATAS_DIR,  'input.xlsx')DbcheckApi(inputexcel).check_order()