代碼筆記,僅供參考,有錯必糾,別買謝謝
限免:6小時
任務
現在要對一堆PDF文檔裏的數據進行數據提取,規整並整理成一個完整的數據表。
PDF裏的數據的格式是這樣的(由於數據不可外泄,所以我進行馬賽克處理,只保留部分數據):
)
上面顯示的表格是我們需要的數據,當然,這個PDF文檔中還有很多我們不需要的干擾數據,所以,我用正則表達式去匹配我需要的數據,並添加了一些條件,禁止干擾數據進入我們最後的數據表中。
這些PDF文檔的名字很規整,方便我們處理,我們可以利用PDF名稱的信息對從提取到的數據進行歸類:
)
這些PDF文件裏的數據分別在2011年、2013年、2015年、2017年和2019年記錄的,所以分別放在下面這5個文件夾裏:
)
因爲我現在還真不知道有什麼直接讀PDF的手段,也不會把PDF轉成txt或者其他格式的技術,所以這裏自己手動複製到Excel中,再轉成csv文件了,轉成csv方便我們讀取(之前倒是試了試用CAJViewer把PDF轉成txt,但是效果不是我想要的那樣):
)
最後,我們需要提取到PDF中result, reaction,DATE of BIRTH,Fn四個變量,並且將數據記錄的年份(比如2019年)和數據分類信息記錄在同一個表中,並導入mysql中,由此,我創建了一個mysql數據表:
CREATE TABLE RunData(id int(11) primary key auto_increment,
PersonDATE date,
RESULT float(4,2),
REACTION float(4,3),
FN int,
CompetitionType varchar(20),
ProjectClass varchar(20),
RunYear YEAR);
python代碼
下面是python代碼:
# -*- coding: utf-8 -*-
import re
import datetime
import os
from pandas import DataFrame
import pandas as pd
import pymysql
class EiDealData:
def __init__(self):
self.path = r"F:\MyStudio\PythonStudio\goatbishop.project01\MyNewWork"
self.RunData = {}
self.countFileNun = 0
self.count1 = 0
self.count2 = 0
self.filter1 = 0
self.filter2 = 0
self.db = pymysql.connect(host = '127.0.0.1',
port = 3306,
user = 'root',
password = '19970928',
database = 'datacup',
charset = 'utf8')
self.cur = self.db.cursor()
def GetFile(self):
fileName = os.listdir(self.path)
self.GetFormatData(fileName)
def GetFormatData(self, fileName):
for item in fileName:
if os.path.isdir(item):
path = self.path + "\\" + item
fileName2 = os.listdir(path)
for file in fileName2:
if file.endswith(".csv"):
self.countFileNun +=1
#print(file)
#判斷後綴
fNameDeal = re.split("\W+", file)
#print(fNameDeal)
sonDictName = "R_"+ "".join(fNameDeal[0:(len(fNameDeal)- 1)])
self.RunData[sonDictName] = {}
self.RunData[sonDictName]["REACTION"] = []
self.RunData[sonDictName]["RESULT"] = []
self.RunData[sonDictName]["DATE"] = []
self.RunData[sonDictName]["FN"] = [] #0爲沒有犯規,1爲犯規
ReadFilePath = path + "\\" + file
with open(ReadFilePath, 'r') as f:
data = f.readline()
#print(data)
while (data):
start_char = data[0]
#flag1 = re.findall(r"\+|\(|\)", data)
#print(flag1)
flag2 = len(re.split("\W", data))
#print(flag2)
if ((ord(start_char) >= 49) & (ord(start_char) <= 57)
& (len(data) >= 16) & (flag2 >= 9)):
self.count1 +=1
#ord("a")爲計算"a"的ASCII碼值
pattern0 = r"\b\d{1,2}\.\d{2}\b"
pattern1 = r"\b\d\.\d{3}\b"
pattern2 = r"\d{1,2} [A-Z][a-zA-Z]{2} \d{2}\b"
result = re.findall(pattern0, data)
reaction = re.findall(pattern1, data)
if result:
result = float(result[0])
else:
temp = re.findall(r"DNF|DQ|DNS", data)
if temp:
if temp[0] == "DNF":
result = -1
elif temp[0] == "DQ":
result = -2
elif temp[0] == "DNS":
result = -3
else:
#print(temp)
#print(data)
result = -4
if reaction:
reaction = float(reaction[0])
else:
data = f.readline()
self.filter1 +=1
continue
try:
date = re.findall(pattern2, data)[0]
except Exception as e:
print("匹配異常,可能是原始數據不規整")
deal_date = "2020-01-01"
else:
deal_date = datetime.datetime.strptime(date,'%d %b %y').strftime('%Y-%m-%d')
#print(date)
fn = re.findall(r"\bF\b|\bF1\b", data)
#print(fn)
if fn:
fn = 1
else:
fn = 0
self.RunData[sonDictName]["RESULT"].append(result)
self.RunData[sonDictName]["REACTION"].append(reaction)
self.RunData[sonDictName]["DATE"].append(deal_date)
self.RunData[sonDictName]["FN"].append(fn)
data = f.readline()
print(item, self.countFileNun)
self.countFileNun = 0
self.writeMysql(item)
self.RunData = {}
def writeCsv(self):
pass
def writeMysql(self, RunYear):
dataDealed = self.RunData
for key in dataDealed:
#print(key)
CompetitionType = re.findall(r"n(Final)|(Round1)|(SemiFinal)|(Preliminary)", key)[0]
CompetitionType = "".join(CompetitionType)
ProjectClass = re.findall(r"\d{3}|hurdles", key)
ProjectClass = "".join(ProjectClass)
#print(len(dataDealed[key]["DATE"]))
self.count2 += len(dataDealed[key]["DATE"])
ProjectClassL = [ProjectClass]*len(dataDealed[key]["DATE"])
#CompetitionType = "".join(cp_list)
#print(CompetitionType)
CompetitionTypeL = [CompetitionType]*len(dataDealed[key]["DATE"])
RunYearL = [RunYear]*len(dataDealed[key]["DATE"])
tempData = zip(dataDealed[key]["DATE"], dataDealed[key]["RESULT"],
dataDealed[key]["REACTION"], dataDealed[key]["FN"],
CompetitionTypeL,ProjectClassL,RunYearL)
for item in tempData:
#print(item)
try:
sql = 'insert into RunData(PersonDATE,RESULT,REACTION,FN,CompetitionType,ProjectClass,RunYear) \
values(%s, %s, %s, %s, %s, %s, %s);'
self.cur.execute(sql, item)
self.db.commit()
print("提交成功...")
except Exception as e:
self.db.rollback()
print("錯誤信息:", e)
def main(self):
self.GetFile()
#self.writeMysql()
self.cur.close()
self.db.close()
if __name__ == "__main__":
eiDeal = EiDealData()
eiDeal.main()
print("進入規則內個數:",eiDeal.count1)
print("實際導入sql個數",eiDeal.count2)
print("過濾掉數據個數:",eiDeal.filter1)
部分輸出:
提交成功...
提交成功...
提交成功...
進入規則內個數: 4051
實際導入sql個數 2463
過濾掉數據個數: 1588
mysql數據庫
現在,我們看看mysql中是否導入2463條數據。
sql語言:
select * from RunData limit 3000;
輸出:
18:17:36 select * from RunData limit 3000 2463 row(s) returned 0.000 sec / 0.015 sec
可以看到,我們已經收集到了2363條數據,現在我們查詢result=-2的記錄:
SELECT * FROM RunData WHERE result=-2;
輸出:
21:31:00 SELECT * FROM RunData WHERE result=-2 LIMIT 0, 1000 19 row(s) returned 0.000 sec / 0.000 sec
可以看到查詢到19條記錄。