|
| 1 | +import requests |
| 2 | +from bs4 import BeautifulSoup |
| 3 | +import re |
| 4 | +import sqlite3 |
| 5 | + |
| 6 | +fdic = {} |
| 7 | +list_a = [] |
| 8 | +list_b = [] |
| 9 | + |
| 10 | +for page in range(1, 55, 1): |
| 11 | + url = "https://fine.fss.or.kr/main/fin_tip/dic/financedic.jsp?page="+str(page) |
| 12 | + print(url) |
| 13 | + |
| 14 | + response = requests.get(url) |
| 15 | + |
| 16 | + if response.status_code == 200: |
| 17 | + html = response.text |
| 18 | + soup = BeautifulSoup(html, 'html.parser') |
| 19 | + # print(soup) |
| 20 | + ul = soup.select_one('ul.dic_result_list') |
| 21 | + sub = ul.select('li > dl > dt') |
| 22 | + con = ul.select('li > dl > dd') |
| 23 | + |
| 24 | + for i in sub: |
| 25 | + tmp = i.get_text().strip() |
| 26 | + tmp = tmp.replace("\r\n\t\t\t\t\t\t\t\t\xa0", "") |
| 27 | + tmp = re.sub('^[0-9]+. ', '', tmp) |
| 28 | + list_a.append(tmp) |
| 29 | + for i in con: |
| 30 | + tmp = i.get_text().strip() |
| 31 | + list_b.append(tmp) |
| 32 | + else: |
| 33 | + print(response.status_code) |
| 34 | + |
| 35 | +for i in range(len(list_a)): |
| 36 | + fdic[i] = [list_a[i], list_b[i]] |
| 37 | +# print(fdic) |
| 38 | + |
| 39 | +conn = sqlite3.connect("fdic1.db") |
| 40 | +cur = conn.cursor() |
| 41 | +try: |
| 42 | + conn.execute("create table fss_dic(id integer, name text, content text)") |
| 43 | +except Exception as e: |
| 44 | + print(e) |
| 45 | + |
| 46 | +for i in fdic: |
| 47 | + name = fdic[i][0] |
| 48 | + content = fdic[i][1] |
| 49 | + |
| 50 | + sql = "insert into fss_dic values (?, ?, ?)" |
| 51 | + cur.execute(sql, (i, name, content)) |
| 52 | + |
| 53 | +conn.commit() |
| 54 | +conn.close() |
0 commit comments