|
@@ -46,8 +46,10 @@ def insert_batch_data(conn, params):
|
|
|
"""
|
|
|
执行批量插入数据
|
|
|
"""
|
|
|
- query = """INSERT IGNORE INTO yusuan_analysis_liantong (mongoid, area, city,district, projectname, publish_org, procure_content,kpi,budget,institution,score, error_type, create_time)
|
|
|
- VALUES (%s, %s, %s, %s, %s, %s, %s, %s,%s,%s,%s,%s,%s )"""
|
|
|
+ query = """INSERT IGNORE INTO zxz_analysis_liantong (mongoid, area, city,district, project_name, total_investment, project_domain,project_owner,start_date,end_date,
|
|
|
+ operation_start_date,operation_end_date,source_income,construction_content,remarks,cost_income_percent,coverage_multiple,competent_department,
|
|
|
+ score, error_type, create_time)
|
|
|
+ VALUES (%s, %s, %s, %s, %s, %s, %s, %s,%s,%s,%s,%s,%s,%s,%s,%s,%s,%s,%s,%s,%s )"""
|
|
|
MysqlUtil.insert_data(conn, query, params)
|
|
|
|
|
|
|
|
@@ -65,7 +67,7 @@ def has_non_empty_qa(data):
|
|
|
|
|
|
def insert_bid_statistics (col,conn,query,batch_id):
|
|
|
#定义来源 1标讯简版2拟在建3新闻4预算5专项债
|
|
|
- data_source =4
|
|
|
+ data_source =5
|
|
|
# 使用聚合管道进行多条件统计
|
|
|
pipeline = [
|
|
|
{"$match": query},
|
|
@@ -79,7 +81,7 @@ def insert_bid_statistics (col,conn,query,batch_id):
|
|
|
# 提取统计结果
|
|
|
count_total = result["总量"][0]["count"] if result["总量"] else 0
|
|
|
|
|
|
- sql_query = """INSERT IGNORE INTO bid_statistics_liantong (yusuan_count, batch_id,data_source)
|
|
|
+ sql_query = """INSERT IGNORE INTO bid_statistics_liantong (zxz_count, batch_id,data_source)
|
|
|
VALUES ( %s, %s ,%s)"""
|
|
|
params = (count_total, batch_id, data_source)
|
|
|
MysqlUtil.insert_data(conn, sql_query, params)
|
|
@@ -98,7 +100,7 @@ def batch_load_data():
|
|
|
end_date = int(datetime.strptime(f"{today_date} 00:00:00", "%Y-%m-%d %H:%M:%S").timestamp())
|
|
|
# print("end_date", end_date)
|
|
|
# 规则查询,根据必要条件 公司名称(用户ID)、版本号
|
|
|
- rules_id = get_rule("中国联通-预算", "v1.4.3")
|
|
|
+ rules_id = get_rule("中国联通-专项债", "v1.4.2")
|
|
|
print(rules_id)
|
|
|
# 初始化mysql
|
|
|
conn = MysqlUtil.connect_to_mysql(host='172.20.45.129', port='4000', user='root', password='=PDT49#80Z!RVv52_z',database='quality')
|
|
@@ -109,7 +111,7 @@ def batch_load_data():
|
|
|
query = {
|
|
|
"_id": {"$gte": max_id},
|
|
|
# "_id": max_id,
|
|
|
- "createtime": {"$gte": start_date, "$lte": end_date}
|
|
|
+ # "createtime": {"$gte": start_date, "$lte": end_date}
|
|
|
}
|
|
|
|
|
|
mongo_client = MongoClient('mongodb://127.0.0.1:27087/', unicode_decode_error_handler="ignore",directConnection=True) # 修改为你的连接地址
|
|
@@ -140,17 +142,28 @@ def batch_load_data():
|
|
|
area = item.get("area", "")
|
|
|
city = item.get("city", "")
|
|
|
district = item.get("district", "")
|
|
|
- projectname = item.get("projectname", "")
|
|
|
- publish_org = item.get("publish_org", "")
|
|
|
- procure_content = item.get("procure_content", "")
|
|
|
- kpi = item.get("kpi", "")
|
|
|
- budget = item.get("budget", "")
|
|
|
- institution = item.get("institution", "")
|
|
|
+ project_name = item.get("project_name", "")
|
|
|
+ total_investment = item.get("total_investment", "")
|
|
|
+ project_domain = item.get("project_domain", "")
|
|
|
+ project_owner = item.get("project_owner", "")
|
|
|
+ start_date = item.get("start_date", "")
|
|
|
+ end_date = item.get("end_date", "")
|
|
|
+ operation_start_date = item.get("operation_start_date", "")
|
|
|
+ operation_end_date = item.get("operation_end_date", "")
|
|
|
+ source_income = item.get("source_income", "")
|
|
|
+ construction_content = item.get("construction_content", "")
|
|
|
+ remarks = item.get("remarks", "")
|
|
|
+ cost_income_percent = item.get("cost_income_percent", "")
|
|
|
+ coverage_multiple = item.get("coverage_multiple", "")
|
|
|
+ competent_department = item.get("competent_department", "")
|
|
|
+
|
|
|
score = data.get("score", "")
|
|
|
error_type_data = json.dumps(data)
|
|
|
create_time = today_date
|
|
|
|
|
|
- params = (item["_id"], area, city, district,projectname, publish_org, procure_content,kpi,budget,institution,score, error_type_data,create_time)
|
|
|
+ params = (item["_id"], area, city, district,project_name, total_investment, project_domain,project_owner,start_date,end_date,
|
|
|
+ operation_start_date,operation_end_date,source_income,construction_content,remarks,cost_income_percent,coverage_multiple,competent_department,
|
|
|
+ score, error_type_data,create_time)
|
|
|
insert_batch_data(conn, params)
|
|
|
|
|
|
print("---- 数据处理完成 ----")
|