My objective is to convert an xls file to xlsx file. The xls file which I am trying to convert is actually an html file containing tables (This xls file is obtained as a result of a query from jira). To facilitate the conversion I have created a file handler and then given that file handler to a beautiful soup and have extracted the table on interest and this extracted table is converted to a string and given to pandas dataframe for further processing.
This works fine but when the file size is large say around 80 MB it takes a large amount of time to process. How do I overcome this?
import bs4, os
import pandas as pd
print('Begin')
fileName = 'TestSample.xls'
fileHandler=open(fileName, encoding='utf-8')
soup = bs4.BeautifulSoup(fileHandler,'html.parser')
tbl = soup.find_all('table', id='issuetable')
df=pd.read_html(str(tbl))
df[0].to_excel("restult.xlsx", index=False)
print('Completed')
There is no good way for large files, but you can try different ways.
from simplified_scrapy import SimplifiedDoc
print('Begin')
fileName = 'TestSample.xls'
html=open(fileName, encoding='utf-8').read()
doc = SimplifiedDoc(html)
start = 0 # If a string can uniquely mark the starting position of data, the performance will be better
tbl = doc.getElement('table', attr='id',value='issuetable', start=start)
print(tbl.outerHtml)
Or block read
f=open(fileName, encoding='utf-8')
html = ''
start = '' # Start of block
end = '' # End of block
for line in f.readlines():
if not html:
html+=line
if line.find(end)>=0:
break
elif line.find(start)>=0:
html = line
if line.find(end)>=0:
break
doc = SimplifiedDoc(html)
tbl = doc.getElement('table', attr='id',value='issuetable')
print(tbl.outerHtml)
Related
I am attempting to scrape the artists' Spotify streaming rankings from Kworb.net into a CSV file and I've nearly succeeded except I'm running into a weird issue.
The code below successfully scrapes all 10,000 of the listed artists into the console:
import requests
from bs4 import BeautifulSoup
import csv
URL = "https://kworb.net/spotify/artists.html"
result = requests.get(URL)
src = result.content
soup = BeautifulSoup(src, 'html.parser')
table = soup.find('table', id="spotifyartistindex")
header_tags = table.find_all('th')
headers = [header.text.strip() for header in header_tags]
rows = []
data_rows = table.find_all('tr')
for row in data_rows:
value = row.find_all('td')
beautified_value = [dp.text.strip() for dp in value]
print(beautified_value)
if len(beautified_value) == 0:
continue
rows.append(beautified_value)
The issue arises when I use the following code to save the output to a CSV file:
with open('artist_rankings.csv', 'w', newline="") as output:
writer = csv.writer(output)
writer.writerow(headers)
writer.writerows(rows)
For whatever reason, only 738 of the artists are saved to the file. Does anyone know what could be causing this?
Thanks so much for any help!
As an alternative approach, you might want to make your life easier next time and use pandas.
Here's how:
import requests
import pandas as pd
source = requests.get("https://kworb.net/spotify/artists.html")
df = pd.concat(pd.read_html(source.text, flavor="bs4"))
df.to_csv("artists.csv", index=False)
This outputs a .csv file with 10,000 artists.
I would like to write new data to the beginning of my text file, with the previous data shifting down 1 line each time new data is imported, I would like everything to be organized, but every time I import something gets deleted.
Code:
import requests
from bs4 import BeautifulSoup
from datetime import datetime
response = requests.get('https://www.lotteryusa.com/michigan/lucky-4-life/')
soup = BeautifulSoup(response.text, 'html.parser')
date = soup.find(class_='date')
results = soup.find(class_='draw-result list-unstyled list-inline')
d = datetime.strptime(date.time['datetime'], '%Y-%m-%d')
Lucky = (d.strftime("%m%d%Y")+(',')+results.get_text()[:-20].strip().replace('\n',','))
print(Lucky)
with open("webscraper2noteppad++", "r+") as f:
file = f.readlines()
f.seek(0,0)
f.write(Lucky)
Also tried doing this
with open("webscraper2noteppad++", "r+") as f:
file = f.read()
f.seek(0,0)
f.write(Lucky + '\n')
but I have to put 10 lines between the already existing data, and the new data. So it can be can be imported on top without deleting.
You can first read the content of your file, the append it to the new data and then write everything to the file:
with open("webscraper2noteppad++", "r") as f:
data = f.read()
with open("webscraper2noteppad++", "w") as f:
f.write('{}{}{}'.format(lucky, '\n' if data else '', data))
I have a bunch of web text that I'd like to scrape and export to a csv file. The problem is that the text is split over multiple lines on the website and that's how beautifulsoup reads it. When I export to csv, all the text goes into one cell but the cell has multiple lines of text. When I try to read the csv into another program, it interprets the multiple lines in a way that yields a nonsensical dataset. The question is, how do I put all the text into a single line after I pull it with beautifulsoup but before I export to csv?
Here's a simple working example demonstrating the problem of multiple lines (in fact, the first few lines in the resulting csv are blank, so at first glance it may look empty):
import csv
import requests
from bs4 import BeautifulSoup
def main():
r = requests.get("https://www.econometricsociety.org/publications/econometrica/2017/03/01/search-yield")
soup = BeautifulSoup(r.text,"html.parser")
with open('Temp.csv', 'w', encoding='utf8', newline='') as f:
writer = csv.writer(f,delimiter=",")
abstract=soup.find("article").text
writer.writerow([abstract])
if __name__ == '__main__':
main()
UPDATE: there have been some good suggestions, but it's still not working. The following code still produces a csv file with line breaks in a cell:
import csv
import requests
from bs4 import BeautifulSoup
with open('Temp.csv', 'w', encoding='utf8', newline='') as f:
writer = csv.writer(f,delimiter=',')
r = requests.get("https://www.econometricsociety.org/publications/econometrica/2017/03/01/search-yield")
soup = BeautifulSoup(r.text,'lxml')
find_article = soup.find('article')
find_2para = find_article.p.find_next_sibling("p")
find_largetxt = find_article.p.find_next_sibling("p").nextSibling
writer.writerow([find_2para,find_largetxt])
Here's another attempt based on a different suggestion. This one also ends up producing a line break in the csv file:
import csv
import requests
from bs4 import BeautifulSoup
def main():
r = requests.get("https://www.econometricsociety.org/publications/econometrica/2017/03/01/search-yield")
soup = BeautifulSoup(r.text,"html.parser")
with open('Temp.csv', 'w', encoding='utf8', newline='') as f:
writer = csv.writer(f,delimiter=",")
abstract=soup.find("article").get_text(separator=" ", strip=True)
writer.writerow([abstract])
if __name__ == '__main__':
main()
Change your abstract = ... line into:
abstract = soup.find("article").get_text(separator=" ", strip=True)
It'll separate each line using the separator parameter (in this case It'll separate the strings with an empty space.
The solution that ended up working for me is pretty simple:
abstract=soup.find("article").text.replace("\t", "").replace("\r", "").replace("\n", "")
That gets rid of all line breaks.
r = requests.get("https://www.econometricsociety.org/publications/econometrica/2017/03/01/search-yield")
soup = BeautifulSoup(r.text,'lxml') # I prefer using xml parser
find_article = soup.find('article')
# Next line how to find The title in this case: Econometrica: Mar 2017, Volume 85, Issue 2
find_title = find_article.h3
# find search yeild
find_yeild = find_article.h1
#first_paragraph example : DOI: 10.3982/ECTA14057 p. 351-378
find_1para = find_article.p
#second p example : David Martinez‐Miera, Rafael Repullo
find_2para = find_article.p.find_next_sibling("p")
#find the large text area using e.g. 'We present a model of the relationship bet...'
find_largetxt = find_article.p.find_next_sibling("p").nextSibling
I used a variety of methods of getting to the text area you wish just for the purpose of education(you can use .text on each of these to get the text without tags or you can use Zroq's method.
But you can write each one of these into the file by doing for example
writer.writerow(find_title.text)
Why do I only get the results form the last url?
The idea is that I get a list of results of both urls.
Also, with the printing in csv I get eacht time an empty row. How do I remove this row?
import csv
import requests
from lxml import html
import urllib
TV_category = ["_108-tot-127-cm-43-tot-50-,98952,501090","_128-tot-150-cm-51-tot-59-,98952,501091"]
url_pattern = 'http://www.mediamarkt.be/mcs/productlist/{}.html?langId=-17'
for item in TV_category:
url = url_pattern.format(item)
page = requests.get(url)
tree = html.fromstring(page.content)
outfile = open("./tv_test1.csv", "wb")
writer = csv.writer(outfile)
rows = tree.xpath('//*[#id="category"]/ul[2]/li')
for row in rows:
price = row.xpath('normalize-space(div/aside[2]/div[1]/div[1]/div/text())')
product_ref = row.xpath('normalize-space(div/div/h2/a/text())')
writer.writerow([product_ref,price])
As I explained in the question's comments, you need to put the second for loop inside (at the end) the first one. Otherwise, only the last rows results will be saved/written to the CSV-format file.
You don't need to open the file in each loop (a with statement will close it automagically). It is, as well, important to highlight that if you open a file with write flags it will overwrite, and if it's inside a loop it will overwrite each time it's opened.
I'd refactor your code as follows:
import csv
import requests
from lxml import html
import urllib
TV_category = ["_108-tot-127-cm-43-tot-50-,98952,501090","_128-tot-150-cm-51-tot-59-,98952,501091"]
url_pattern = 'http://www.mediamarkt.be/mcs/productlist/{}.html?langId=-17'
with open("./tv_test1.csv", "wb") as outfile:
writer = csv.writer(outfile)
for item in TV_category:
url = url_pattern.format(item)
page = requests.get(url)
tree = html.fromstring(page.content)
rows = tree.xpath('//*[#id="category"]/ul[2]/li')
for row in rows:
price = row.xpath('normalize-space(div/aside[2]/div[1]/div[1]/div/text())')
product_ref = row.xpath('normalize-space(div/div/h2/a/text())')
writer.writerow([product_ref,price])
I'm trying to extract the data on the crime rate across states from
this webpage, link to web page
http://www.disastercenter.com/crime/uscrime.htm
I am able to get this into text file. But I would like to get the
response in Json format. How can I do this in python.
Here is my code:
import urllib
import re
from bs4 import BeautifulSoup
link = "http://www.disastercenter.com/crime/uscrime.htm"
f = urllib.urlopen(link)
myfile = f.read()
soup = BeautifulSoup(myfile)
soup1=soup.find('table', width="100%")
soup3=str(soup1)
result = re.sub("<.*?>", "", soup3)
print(result)
output=open("output.txt","w")
output.write(result)
output.close()
The following code will get the data from the two tables and output all of it as a json formatted string.
Working Example (Python 2.7.9):
from lxml import html
import requests
import re as regular_expression
import json
page = requests.get("http://www.disastercenter.com/crime/uscrime.htm")
tree = html.fromstring(page.text)
tables = [tree.xpath('//table/tbody/tr[2]/td/center/center/font/table/tbody'),
tree.xpath('//table/tbody/tr[5]/td/center/center/font/table/tbody')]
tabs = []
for table in tables:
tab = []
for row in table:
for col in row:
var = col.text_content()
var = var.strip().replace(" ", "")
var = var.split('\n')
if regular_expression.match('^\d{4}$', var[0].strip()):
tab_row = {}
tab_row["Year"] = var[0].strip()
tab_row["Population"] = var[1].strip()
tab_row["Total"] = var[2].strip()
tab_row["Violent"] = var[3].strip()
tab_row["Property"] = var[4].strip()
tab_row["Murder"] = var[5].strip()
tab_row["Forcible_Rape"] = var[6].strip()
tab_row["Robbery"] = var[7].strip()
tab_row["Aggravated_Assault"] = var[8].strip()
tab_row["Burglary"] = var[9].strip()
tab_row["Larceny_Theft"] = var[10].strip()
tab_row["Vehicle_Theft"] = var[11].strip()
tab.append(tab_row)
tabs.append(tab)
json_data = json.dumps(tabs)
output = open("output.txt", "w")
output.write(json_data)
output.close()
This might be what you want, if you can use the requests and lxml modules. The data structure presented here is very simple, adjust this to your needs.
First, get a response from your requested URL and parse the result into an HTML tree:
import requests
from lxml import etree
import json
response = requests.get("http://www.disastercenter.com/crime/uscrime.htm")
tree = etree.HTML(response.text)
Assuming you want to extract both tables, create this XPath and unpack the results. totals is "Number of Crimes" and rates is "Rate of Crime per 100,000 People":
xpath = './/table[#width="100%"][#style="background-color: rgb(255, 255, 255);"]//tbody'
totals, rates = tree.findall(xpath)
Extract the raw data (td.find('./') means first child item, whatever tag it has) and clean the strings (r'' raw strings are needed for Python 2.x):
raw_data = []
for tbody in totals, rates:
rows = []
for tr in tbody.getchildren():
row = []
for td in tr.getchildren():
child = td.find('./')
if child is not None and child.tag != 'br':
row.append(child.text.strip(r'\xa0').strip(r'\n').strip())
else:
row.append('')
rows.append(row)
raw_data.append(rows)
Zip together the table headers in the first two rows, then delete the redundant rows, seen as the 11th & 12th steps in slice notation:
data = {}
data['tags'] = [tag0 + tag1 for tag0, tag1 in zip(raw_data[0][0], raw_data[0][1])]
for raw in raw_data:
del raw[::12]
del raw[::11]
Store the rest of the raw data and create a JSON file (optional: eliminate whitespace with separators=(',', ':')):
data['totals'], data['rates'] = raw_data[0], raw_data[1]
with open('data.json', 'w') as f:
json.dump(data, f, separators=(',', ':'))