def fetch_dupont_data(stocklist, start_year=2007, current_year=2020, current_quarter=4):
bs.login()
db_conn = create_engine(common.db_path_sqlalchemy)
for code in stocklist["code"]:
dupont_list = []
if code.startswith("sh.6") | code.startswith("sz.00") | code.startswith("sz.300"):
for year in range(start_year, current_year + 1):
start_quarter = 1
if year < current_year:
end_quarter = 4
else:
end_quarter = current_quarter - 1
for quarter in range(start_quarter, end_quarter + 1):
# 查询季频估值指标盈利能力
print(code + ' dupont: ' + year.__str__() + 'Q' + quarter.__str__())
rs_dupont = bs.query_dupont_data(code=code, year=year, quarter=quarter)
while (rs_dupont.error_code == '0') & rs_dupont.next():
dupont_list.append([year, quarter] + rs_dupont.get_row_data())
if len(dupont_list) > 0:
db_conn.execute(r'''INSERT OR REPLACE INTO stock_dupont_data VALUES (?,?,?,?,?,?,?,?,?,?,?,?,?)''',
dupont_list)
print(code + ' fetch dupont data finish and write database')
bs.logout()
def get_dupont_data(stock_number,stock_name,year,quarter):
print('==========================================================')
print("开始进行: "+stock_name+"("+stock_number+")"+"的数据处理")
print("尝试登陆baostock")
#####login#####
lg=bs.login(user_id="anonymous",password="******")
if(lg.error_code == '0'):
print("登陆成功")
else:
print("登录失败")
#####get stock data#####
rs=bs.query_dupont_data(code=stock_number,year=year,quarter=quarter)
print('请求历史数据返回代码:'+rs.error_code)
print('请求历史数据返回信息:'+rs.error_msg)
data_list=[]
while(rs.error_code=='0')&rs.next():
data_list.append(rs.get_row_data())
result=pd.DataFrame(data_list,columns=rs.fields)
bs.logout()
print(stock_name+"("+stock_number+")"+"的数据处理完成")
print('==========================================================')
return result
def queryDubpontByCode(code,year,quater):
dupont_list = []
rs_dupont = bs.query_dupont_data(code=code, year=year, quarter=quater)
while (rs_dupont.error_code == '0') & rs_dupont.next():
dupont_list.append(rs_dupont.get_row_data())
result_profit = pd.DataFrame(dupont_list, columns=rs_dupont.fields)
return result_profit
def download(stockCode, yearList: list, fileName):
#### 登陆系统 ####
lg = bs.login()
# 显示登陆返回信息
print('login respond error_code:' + lg.error_code)
print('login respond error_msg:' + lg.error_msg)
# 查询季频估值指标盈利能力
operation_list = []
for year in yearList:
for quarter in [1, 2, 3, 4]:
rs_dupont = bs.query_dupont_data(code=stockCode,
year=year,
quarter=quarter)
# rs_dupont = bs.query_dupont_data(code="sh.600000", year=2017, quarter=2)
# rs_growth = bs.query_growth_data(code="sh.600000", year=2017, quarter=2)
while (rs_dupont.error_code == '0') & rs_dupont.next():
operation_list.append(rs_dupont.get_row_data())
saveToCsv(operation_list, fileName)
#### 登出系统 ####
bs.logout()
def getROEQuery(stock, year, quarter):
dupont_list = []
rs_dupont = bs.query_dupont_data(stock['code'], year, quarter)
while (rs_dupont.error_code == '0') & rs_dupont.next():
dupont_list.append(rs_dupont.get_row_data())
result_profit = pd.DataFrame(dupont_list,
columns=rs_dupont.fields)
return result_profit
def getquarterdata(companycode, year, quarter):
profit_list = []
rs_profit = bs.query_profit_data(code=companycode,
year=year,
quarter=quarter)
while (rs_profit.error_code == '0') & rs_profit.next():
profit_list.append(rs_profit.get_row_data())
if len(profit_list) == 0:
return [], []
# 营运能力
operation_list = []
rs_operation = bs.query_operation_data(code=companycode,
year=year,
quarter=quarter)
while (rs_operation.error_code == '0') & rs_operation.next():
operation_list.append(rs_operation.get_row_data())
# 成长能力
growth_list = []
rs_growth = bs.query_growth_data(code=companycode,
year=year,
quarter=quarter)
while (rs_growth.error_code == '0') & rs_growth.next():
growth_list.append(rs_growth.get_row_data())
# 偿债能力
balance_list = []
rs_balance = bs.query_balance_data(code=companycode,
year=year,
quarter=quarter)
while (rs_balance.error_code == '0') & rs_balance.next():
balance_list.append(rs_balance.get_row_data())
# 季频现金流量
cash_flow_list = []
rs_cash_flow = bs.query_cash_flow_data(code=companycode,
year=year,
quarter=quarter)
while (rs_cash_flow.error_code == '0') & rs_cash_flow.next():
cash_flow_list.append(rs_cash_flow.get_row_data())
# 查询杜邦指数
dupont_list = []
rs_dupont = bs.query_dupont_data(code=companycode,
year=year,
quarter=quarter)
while (rs_dupont.error_code == '0') & rs_dupont.next():
dupont_list.append(rs_dupont.get_row_data())
if len(profit_list) * len(operation_list) * len(growth_list) * len(
balance_list) * len(cash_flow_list) * len(dupont_list) == 0:
return [], []
datalist = [companycode + "-" + year + "-" + quarter]
datalist += profit_list[0] + operation_list[0] + growth_list[
0] + balance_list[0] + cash_flow_list[0] + dupont_list[0]
fieldslist = ["quaryContent"]
fieldslist += rs_profit.fields + rs_operation.fields + rs_growth.fields + rs_balance.fields + rs_cash_flow.fields + rs_dupont.fields
return datalist, fieldslist
def getROE(code="SH.600000", year=2015):
rs_dupont = bs.query_dupont_data(code=code, year=year, quarter=4)
re = 0
try:
re = rs_dupont.get_row_data()[3]
re = round(float(re) * 100, 2)
except:
pass
return re
def computeROE(code, year, quarter):
# 查询杜邦指数
dupont_list = []
rs_dupont = bs.query_dupont_data(code, year, quarter)
while (rs_dupont.error_code == '0') & rs_dupont.next():
dupont_list.append(rs_dupont.get_row_data())
result_profit = pd.DataFrame(dupont_list, columns=rs_dupont.fields)
# 打印输出
return result_profit
def dupont(code_n, year_n, quarter_n):
dupont_list = []
rs_dupont = bs.query_dupont_data(code=code_n,
year=year_n,
quarter=quarter_n)
while (rs_dupont.error_code == '0') & rs_dupont.next():
dupont_list.append(rs_dupont.get_row_data())
result_profit = pd.DataFrame(dupont_list, columns=rs_dupont.fields)
# 打印输出
return result_profit
def get_dupont(code):
dupont_list = []
for year in range(2000, 2020):
for quarter in [4]:
rs_dupont = bs.query_dupont_data(code=code,
year=year,
quarter=quarter)
while (rs_dupont.error_code == '0') & rs_dupont.next():
dupont_list.append(rs_dupont.get_row_data())
result_profit = pd.DataFrame(dupont_list, columns=rs_dupont.fields)
return result_profit
def query_dupont_data(self, code, year = None, quarter = None):
'''
code：股票代码，sh或sz.+6位数字代码，或者指数代码，如：sh.601398。sh：上海；sz：深圳。此参数不可为空；
year：统计年份，为空时默认当前年；
quarter：统计季度，为空时默认当前季度。不为空时只有4个取值：1，2，3，4。
'''
dupont_list = []
rs_dupont = bs.query_dupont_data(code=code, year=year, quarter=quarter)
self.log('query_dupont_data respond error_msg:', rs_dupont)
while (rs_dupont.error_code == '0') & rs_dupont.next():
dupont_list.append(rs_dupont.get_row_data())
result_profit = pd.DataFrame(dupont_list, columns=rs_dupont.fields)
return result_profit
def query_dupont_data(self):
"""
季频杜邦指数：query_dupont_data()
方法说明：查询季频杜邦指数信息，可以通过参数设置获取对应年份、季度数据，提供2007年至今数据。
返回类型：pandas的DataFrame类型。
"""
all_stock = pd.read_csv(self.path + "all_stock.csv")
all_stock.drop_duplicates(subset=['code'], keep='first', inplace=True)
end_year = int(self.end_date[0:4])
quarter_list = [1, 2, 3, 4]
# 登陆系统
lg = bs.login()
# 显示登陆返回信息
#print('login respond error_code:'+lg.error_code)
#print('login respond error_msg:'+lg.error_msg)
# 查询杜邦指数
result_all = pd.DataFrame([])
for code in all_stock.code:
# 查询季频估值指标盈利能力
year = int(self.start_date[0:4])
while (year <= end_year):
for quarter in quarter_list:
dupont_list = []
rs_dupont = bs.query_dupont_data(code=code,
year=year,
quarter=quarter)
while (rs_dupont.error_code == '0') & rs_dupont.next():
dupont_list.append(rs_dupont.get_row_data())
result_profit = pd.DataFrame(dupont_list,
columns=rs_dupont.fields)
result_all = result_all.append(result_profit)
year = year + 1
# 打印输出
#print(result_profit)
# 结果集输出到csv文件
dupont_data = pd.read_csv(self.path + "dupont_data.csv")
result_all = result_all.append(dupont_data)
result_all.drop_duplicates(subset=['code', 'pubDate', 'statDate'],
keep='first',
inplace=True)
result_all.to_csv(self.path + "dupont_data.csv",
encoding="gbk",
index=False)
# 登出系统
bs.logout()
return
def dupont(code, year, quarter):
filename = "data/dupont_{}_{}_{}.csv".format(code, year, quarter)
if os.path.exists(filename):
# get from local
result_profile = pd.read_csv(filename)
else:
# get from remote
dupont_list = []
rs_dupont = bs.query_dupont_data(code=code, year=year, quarter=quarter)
while (rs_dupont.error_code == '0') & rs_dupont.next():
dupont_list.append(rs_dupont.get_row_data())
result_profile = pd.DataFrame(dupont_list, columns=rs_dupont.fields)
result_profile.to_csv(filename, encoding="utf-8", index=False)
return result_profile
def getFinDataFromBS(self, codes: list, year: int, quarter: int):
#盈利能力
bs.login()
for code in codes:
ay = code.split(".")
exchg = ay[0]
rawcode = ay[1]
bscode = ''
if exchg == 'SSE':
bscode = 'sh.' + rawcode
else:
bscode = 'sz.' + rawcode
fdata = dict()
fdata['code'] = code
rs = bs.query_profit_data(bscode, year, quarter)
fdata = recordToDict(rs, fdata)
rs = bs.query_operation_data(bscode, year, quarter)
fdata = recordToDict(rs, fdata)
rs = bs.query_growth_data(bscode, year, quarter)
fdata = recordToDict(rs, fdata)
rs = bs.query_balance_data(bscode, year, quarter)
fdata = recordToDict(rs, fdata)
rs = bs.query_cash_flow_data(bscode, year, quarter)
fdata = recordToDict(rs, fdata)
rs = bs.query_dupont_data(bscode, year, quarter)
fdata = recordToDict(rs, fdata)
# rs = bs.query_performance_express_report(bscode, year, quarter)
# fdata = recordToDict(rs, fdata)
# rs = bs.query_forcast_report(bscode, year, quarter)
# fdata = recordToDict(rs, fdata)
bs.logout()
return fdata
def get_dupont_data_year(stock_number,stock_name,start_year,end_year):
print('==========================================================')
print("开始进行: "+stock_name+"("+stock_number+")"+"的数据处理")
print("尝试登陆baostock")
data_list=[]
lg=bs.login(user_id="anonymous",password="******")
if(lg.error_code == '0'):
print("登陆成功")
else:
print("登录失败")
for y in range(int(start_year),int(end_year)+1):
for q in range(1,5):
#####get stock data#####
rs=bs.query_dupont_data(code=stock_number,year=y,quarter=q)
while(rs.error_code=='0')&rs.next():
data_list.append(rs.get_row_data())
result=pd.DataFrame(data_list,columns=rs.fields)
bs.logout()
print(stock_name+"("+stock_number+")"+"的数据处理完成")
print('==========================================================')
return result
def GetCurrentStockFundamentals(self, *args, **kwargs):
post_data_code = CheckKeyOrAbort(kStockCodeKey, kwargs)
profit_data_handler = bs.query_profit_data(post_data_code)
profit_data = profit_data_handler.get_data()
operation_data_handler = bs.query_operation_data(post_data_code)
operation_data = operation_data_handler.get_data()
growth_data_handler = bs.query_growth_data(post_data_code)
growth_data = growth_data_handler.get_data()
balance_data_handler = bs.query_balance_data(post_data_code)
balance_data = balance_data_handler.get_data()
cash_flow_data_handler = bs.query_balance_data(post_data_code)
cash_flow_data = cash_flow_data_handler.get_data()
dupont_data_handler = bs.query_dupont_data(post_data_code)
dupont_data = dupont_data_handler.get_data()
performance_data_handler = bs.query_performance_express_report(
post_data_code)
performance_data = performance_data_handler.get_data()
forecast_data_handler = bs.query_forecast_report(post_data_code)
forecast_data = forecast_data_handler.get_data()
return profit_data, operation_data, growth_data, balance_data, cash_flow_data, dupont_data, performance_data, forecast_data
def query_dupont_data(code, year=None, quarter=None):
"""
季频杜邦指数
方法说明：通过API接口获取季频杜邦指数信息，可以通过参数设置获取对应年份、季度数据，提供2007年至今数据。
返回类型：pandas的DataFrame类型。
参数含义：
code：股票代码，sh或sz.+6位数字代码，或者指数代码，如：sh.601398。sh：上海；sz：深圳。此参数不可为空；
year：统计年份，为空时默认当前年；
quarter：统计季度，为空时默认当前季度。不为空时只有4个取值：1，2，3，4。
"""
lg = bs.login()
if lg.error_code != '0':
logger.error('login respond error_msg:' + lg.error_msg)
dupont_list = []
rs_dupont = bs.query_dupont_data(code=code, year=year, quarter=quarter)
while (rs_dupont.error_code == '0') & rs_dupont.next():
dupont_list.append(rs_dupont.get_row_data())
result_profit = pd.DataFrame(dupont_list, columns=rs_dupont.fields)
bs.logout()
return result_profit
import baostock as bs
import pandas as pd
# 登陆系统
lg = bs.login()
# 显示登陆返回信息
print('login respond error_code:' + lg.error_code)
print('login respond error_msg:' + lg.error_msg)
# 查询杜邦指数
dupont_list = []
rs_dupont = bs.query_dupont_data(code="sz.000001", year=2019, quarter=2)
while (rs_dupont.error_code == '0') & rs_dupont.next():
dupont_list.append(rs_dupont.get_row_data())
result_profit = pd.DataFrame(dupont_list, columns=rs_dupont.fields)
# 打印输出
print(result_profit)
# 结果集输出到csv文件
result_profit.to_csv("D:\\dupont_data.csv", encoding="gbk", index=False)
# 登出系统
bs.logout()
def getSeasonDataOneCode(i):
import baostock as bs
import pandas as pd
# #### 登陆系统 ####
# lg = bs.login()
profit_list = []
growth_list = []
dupont_list = []
code = codeFormat(i)
for j in range(2015, 2023):
for k in range(1, 5):
# 读取季度盈利
rs_profit = bs.query_profit_data(code=code, year=j, quarter=k)
while (rs_profit.error_code == '0') & rs_profit.next():
profit_list.append(rs_profit.get_row_data())
# 读取季度成长
rs_growth = bs.query_growth_data(code=code, year=j, quarter=k)
while (rs_growth.error_code == '0') & rs_growth.next():
growth_list.append(rs_growth.get_row_data())
# 读取杜邦分析
rs_dupont = bs.query_dupont_data(code=code, year=j, quarter=k)
while (rs_dupont.error_code == '0') & rs_dupont.next():
dupont_list.append(rs_dupont.get_row_data())
# 处理盈利分析
profit_list = profit_list[::-1]
if profit_list != []:
epsTTMArray = [i, "每股收益/epsTTM",
dateSeasonFormat(profit_list[0][2])] # 每股收益
# roeAvgArray = [i, "净资产收益率/roeAvg",profit_list[0][2]] #净资产收益率
for x in profit_list:
x[2] = dateSeasonFormat(x[2])
if x[7] != '':
epsTTMArray.append(round(float(x[7]), 2))
else:
epsTTMArray.append(0)
# roeAvgArray.append(round(float(x[3]),2))
# 处理成长
growth_list = growth_list[::-1]
yoyniArray = [
i, "净利润同比增长率G/yoyni",
dateSeasonFormat(growth_list[0][2])
] # 净利润同比增长率
for x in growth_list:
x[2] = dateSeasonFormat(x[2])
if x[5] == '':
yoyniArray.append('')
else:
yoyniArray.append(round(float(x[5]), 2))
# 处理杜邦分析
dupont_list = dupont_list[::-1]
dupontROEArray = [
i, "净资产收益率/dupontROE",
dateSeasonFormat(dupont_list[0][2])
] # 杜邦表ROE，净资产收益率
dupontNitogrArray = [
i, "销售净利润率/dupontNitogr",
dateSeasonFormat(dupont_list[0][2])
] # 销售净利润率，净利润/营业总收入
dupontAssetTurnArray = [
i, "总资产周转率/dupontAssetTurn",
dateSeasonFormat(dupont_list[0][2])
] # 总资产周转率
dupontAssetStoEquityArray = [
i, "权益乘数/dupontAssetStoEquity",
dateSeasonFormat(dupont_list[0][2])
] # 权益乘数
for x in dupont_list:
x[2] = dateSeasonFormat(x[2])
if x[3] == '':
dupontROEArray.append('')
else:
dupontROEArray.append(round(float(x[3]), 3))
if x[7] == '':
dupontNitogrArray.append('')
else:
dupontNitogrArray.append(round(float(x[7]), 2))
if x[5] == '':
dupontAssetTurnArray.append('')
else:
dupontAssetTurnArray.append(round(float(x[5]), 2))
if x[4] == '':
dupontAssetStoEquityArray.append('')
else:
dupontAssetStoEquityArray.append(round(float(x[4]), 2))
# 把数据聚合
# array.append(epsTTMArray)
# array.append(yoyniArray)
# array.append(dupontROEArray)
# array.append(dupontNitogrArray)
# array.append(dupontAssetTurnArray)
# array.append(dupontAssetStoEquityArray)
# bs.logout()
return [
epsTTMArray, yoyniArray, dupontROEArray, dupontNitogrArray,
dupontAssetTurnArray, dupontAssetStoEquityArray
]
#### 登出系统 ####
else:
# bs.logout()
return [[], [], [], [], [], []]
import baostock as bs
import pandas as pd
# 登陆系统
lg = bs.login(user_id="anonymous", password="******")
# 显示登陆返回信息
print('login respond error_code:' + lg.error_code)
print('login respond error_msg:' + lg.error_msg)
# 查询杜邦指数
dupont_list = []
rs_dupont = bs.query_dupont_data(code="sh.600000", year=2017, quarter=2)
while (rs_dupont.error_code == '0') & rs_dupont.next():
dupont_list.append(rs_dupont.get_row_data())
result_profit = pd.DataFrame(dupont_list, columns=rs_dupont.fields)
# 打印输出
print(result_profit)
# 结果集输出到csv文件
result_profit.to_csv("D:\\dupont_data.csv", encoding="gbk", index=False)
# 登出系统
bs.logout()
def quarter_dupont_data(self, code, year, quarter):
self.getStockInstance()
rs = bs.query_dupont_data(code, year, quarter)
self._test_rs(rs)
return rs
