feat: 10장 조회한 데이터 DB 저장

example/10-stock-data.py

@@ -1,4 +1,5 @@
 import logging
+import os
 import re
 from io import BytesIO
 
@@ -6,6 +7,10 @@ import numpy as np
 import pandas as pd
 import requests as rq
 from bs4 import BeautifulSoup
+import pymysql
+from dotenv import load_dotenv
+
+load_dotenv()
 
 GEN_OTP_URL = 'http://data.krx.co.kr/comm/fileDn/GenerateOTP/generate.cmd'
 DOWN_URL = 'http://data.krx.co.kr/comm/fileDn/download_csv/download.cmd'
@@ -95,25 +100,49 @@ def get_total_stock_data(biz_day):
     # 데이터 정리
     # 종목, 개별 중 한군데만 있는 데이터 삭제(선박펀드, 광물펀드, 해외종목 등)
     diff = list(set(krx_sector['종목명']).symmetric_difference(set(krx_ind['종목명'])))
-    kor_ticket = pd.merge(krx_sector, krx_ind, on=krx_sector.columns.intersection(krx_ind.columns).tolist(), how='outer')
+    kor_ticker = pd.merge(krx_sector, krx_ind, on=krx_sector.columns.intersection(krx_ind.columns).tolist(), how='outer')
     # 일반적인 종목과 SPAC, 우선주, 리츠, 기타 주식을 구분
-    kor_ticket['종목구분'] = np.where(kor_ticket['종목명'].str.contains('스팩|제[0-9]+호'), '스팩',
+    kor_ticker['종목구분'] = np.where(kor_ticker['종목명'].str.contains('스팩|제[0-9]+호'), '스팩',
-            np.where(kor_ticket['종목코드'].str[-1:] != '0', '우선주',
+            np.where(kor_ticker['종목코드'].str[-1:] != '0', '우선주',
-                     np.where(kor_ticket['종목명'].str.endswith('리츠'), '리츠',
+                     np.where(kor_ticker['종목명'].str.endswith('리츠'), '리츠',
-                              np.where(kor_ticket['종목명'].isin(diff), '기타',
+                              np.where(kor_ticker['종목명'].isin(diff), '기타',
                                        '보통주'
             ))))
-    kor_ticket = kor_ticket.reset_index(drop=True)
+    kor_ticker = kor_ticker.reset_index(drop=True)
-    kor_ticket.columns = kor_ticket.columns.str.replace(' ', '')
+    kor_ticker.columns = kor_ticker.columns.str.replace(' ', '')
-    kor_ticket = kor_ticket[['종목코드', '종목명', '시장구분', '종가',
+    kor_ticker = kor_ticker[['종목코드', '종목명', '시장구분', '종가',
                              '시가총액', '기준일', 'EPS', '선행EPS', 'BPS', '주당배당금', '종목구분']]
-    kor_ticket = kor_ticket.replace({np.nan: None})
+    kor_ticker = kor_ticker.replace({np.nan: None})
-    kor_ticket['기준일'] = pd.to_datetime(kor_ticket['기준일'])
+    kor_ticker['기준일'] = pd.to_datetime(kor_ticker['기준일'])
-    return kor_ticket
+    return kor_ticker
+
+
+def save_db(ticker):
+    con = pymysql.connect(user=os.getenv('DB_USER'),
+                          passwd=os.getenv('DB_PW'),
+                          host=os.getenv('DB_HOST'),
+                          port=int(os.getenv('DB_PORT')),
+                          db=os.getenv('DB_DB'),
+                          charset='utf8')
+
+    mycursor = con.cursor()
+    query = f"""
+        insert into kor_ticker (종목코드,종목명,시장구분,종가,시가총액,기준일,EPS,선행EPS,BPS,주당배당금,종목구분)
+        values (%s,%s,%s,%s,%s,%s,%s,%s,%s,%s,%s) as new
+        on duplicate key update
+        종목명=new.종목명,시장구분=new.시장구분,종가=new.종가,시가총액=new.시가총액,EPS=new.EPS,선행EPS=new.선행EPS,
+        BPS=new.BPS,주당배당금=new.주당배당금,종목구분 = new.종목구분;
+    """
+
+    args = ticker.values.tolist()
+
+    mycursor.executemany(query, args)
+    con.commit()
+
+    con.close()
+
 
 if __name__ == '__main__':
     latest_biz_day = get_latest_biz_day()
     data = get_total_stock_data(latest_biz_day)
-    print(data)
+    save_db(data)
-
-