1

I try to flatten some columns in my dataframe, but unfurtunately it does not work. What would be the correct way of doing this?

created_at tweet_hashtag tweet_cashtag
2022-07-23 [{'start': 16, 'end': 27, 'tag': 'blockchain'}, {'start': 28, 'end': 32, 'tag': 'btc'}, {'start': 33, 'end': 37, 'tag': 'eth'}, {'start': 38, 'end': 42, 'tag': 'eth'}] [{'start': 0, 'end': 4, 'tag': 'Act'}, {'start': 7, 'end': 11, 'tag': 'jar'}]
2022-04-24 [{'start': 6, 'end': 7, 'tag': 'chain'}, {'start': 8, 'end': 3, 'tag': 'btc'}, {'start': 3, 'end': 7, 'tag': 'eth'}] [{'start': 4, 'end': 8, 'tag': 'Act'}, {'start': 7, 'end': 9, 'tag': 'aapl'}]

And my preferred result would be:

created_at tweet_hashtag.tag tweet_cashtag.tag
2022-07-23 blockchain, btc, eth,eth Act, jar
2022-04-24 chain, btc, eth Act, aapl

Thanks in advance!

I tried to flatten with this solution, but it does not work: How to apply json_normalize on entire pandas column

tiktak
  • 11
  • 1

1 Answers1

0

you can use:

def get_values(a,b):
    x_values=[]
    for i in range(0,len(a)):
        x_values.append(a[i]['tag'])
    y_values=[]
    for j in range(0,len(b)):
        y_values.append(b[j]['tag'])
    return ','.join(x_values),','.join(y_values)

df[['tweet_hashtag','tweet_cashtag']]=df[['tweet_hashtag','tweet_cashtag']].apply(lambda x: get_values(x['tweet_hashtag'], x['tweet_cashtag']),axis=1)

or:

def get_hashtags(a):
    x_values=[]
    for i in range(0,len(a)):
        x_values.append(a[i]['tag'])
    return ','.join(x_values)
        
def get_cashtags(b):
    y_values=[]
    for i in range(0,len(b)):
        y_values.append(b[i]['tag'])        
    return ','.join(y_values)

df['tweet_hashtag']=df['tweet_hashtag'].apply(lambda x: get_hashtags(x))
df['tweet_cashtag']=df['tweet_cashtag'].apply(lambda x: get_cashtags(x))
print(df)

'''
    created_at  tweet_hashtag           tweet_cashtag
0   2022-07-23  blockchain,btc,eth,eth  Act,jar
1   2022-04-24  chain,btc,eth           Act,aapl
'''
Bushmaster
  • 4,196
  • 3
  • 8
  • 28