Append returned key values from function as new columns to Dataframe

Question

I have a dataframe with a list of urls for which I want to extract a couple of values. The returned key/values should then be added to the original dataframe with the keys as new columns and the respective values.

I thought that this would magically happen with result_type='expand' which it obviously doesn't. When I try

df5["data"] = df5.apply(lambda x: request_function(x['url']),axis=1, result_type='expand')

I end up with my results all in one data column:

[{'title': ['Python Notebooks: Connect to Google Search Console API and Extract Data - Adapt'], 'description': []}]

The result I am aiming for is a Dataframe with the following 3 columns:

| URL|      Title      |  Description|

Here is my code:

import requests
from requests_html import HTMLSession
import pandas as pd
from urllib import parse

ex_dic = {'url': ['https://www.searchenginejournal.com/reorganizing-xml-sitemaps-python/295539/', 'https://searchengineland.com/check-urls-indexed-google-using-python-259773', 'https://adaptpartners.com/technical-seo/python-notebooks-connect-to-google-search-console-api-and-extract-data/']}

df5 = pd.DataFrame(ex_dic)
df5

def request_function(url):
    try:
        found_results = []
        r = session.get(url)
        title = r.html.xpath('//title/text()')
        description = r.html.xpath("//meta[@name='description']/@content")
        found_results.append({ 'title': title, 'description': description})
        return found_results


    except requests.RequestException:
        print("Connectivity error")      
    except (KeyError):
        print("anoter error")

df5.apply(lambda x: request_function(x['url']),axis=1, result_type='expand')

Answer 1

ex_dic should be list of dict, so that you can update the applied attribute.

import requests
from requests_html import HTMLSession
import pandas as pd
from urllib import parse

ex_dic = {'url': ['https://www.searchenginejournal.com/reorganizing-xml-sitemaps-python/295539/', 'https://searchengineland.com/check-urls-indexed-google-using-python-259773', 'https://adaptpartners.com/technical-seo/python-notebooks-connect-to-google-search-console-api-and-extract-data/']}

ex_dic['url'] = [{'url': item} for item in ex_dic['url']]

df5 = pd.DataFrame(ex_dic)
session = HTMLSession()

def request_function(url):
    try:
        print(url)
        r = session.get(url['url'])
        title = r.html.xpath('//title/text()')
        description = r.html.xpath("//meta[@name='description']/@content")
        url.update({ 'title': title, 'description': description})
        return url


    except requests.RequestException:
        print("Connectivity error")      
    except (KeyError):
        print("anoter error")

df6 = df5.apply(lambda x: request_function(x['url']),axis=1, result_type='expand')
print df6

Answer 2

It actually works as you expect, if your function would return just a dictionary, not a list of dictionaries. Further, inside of your keys just provide a string, not a list. Then it works as you expect. See my example code:

import requests
import pandas as pd
from urllib import parse

ex_dic = {'url': ['https://www.searchenginejournal.com/reorganizing-xml-sitemaps-python/295539/', 'https://searchengineland.com/check-urls-indexed-google-using-python-259773', 'https://adaptpartners.com/technical-seo/python-notebooks-connect-to-google-search-console-api-and-extract-data/']}

df5 = pd.DataFrame(ex_dic)
#rint(df5)

def request_function(url):
    return {'title': 'Python Notebooks: Connect to Google Search Console API and Extract Data - Adapt', 
            'description': ''}


df6 = df5.apply(lambda x: request_function(x['url']), axis=1, result_type='expand')
df7 = pd.concat([df5,df6],1)


df7

Gives you this:

You can also just adjust your lambda function:

df6 = df5.apply(lambda x: request_function(x['url'])[0], axis=1, result_type='expand')

But you still need to ensure that the key values are strings, not lists.