如果值采用类似字典的格式，则从熊猫系列中提取数据

Question

I try the solution in Extracting dictionary values from a pandas dataframe But it didn't work.

I have a pandas.core.series.Series with the following general format:

0      {'hashtags': [], 'symbols': [], 'user_mentions...
1      {'hashtags': [], 'symbols': [], 'user_mentions...
2      {'hashtags': [], 'symbols': [], 'user_mentions...
3      {'hashtags': [], 'symbols': [], 'user_mentions...
...

the specific format of each one is similar to the following:

{'hashtags': [],
'symbols': [],
'user_mentions': [{'screen_name': 'jose_m',
'id_str': '132',
'name': 'Jose',
'indices': [0, 10],
'id': 103},
{'screen_name': 'paul',
 'id_str': '243403',
 'name': 'Jorge',
 'indices': [50, 64],
 'id': 2423}],
 'urls': []}

I get that by placing the index zero to the variable entities[0] (Index may change).

I need to extract extract all the screen_name and name inside user_mentions. Thanks :)

Answer 1

Here is an example with apply , for each entities returns a list with a tuple for each user_mention :

def find_user_mention(user_mention):
    return (user_mention['screen_name'], user_mention['name'])

df['entities'].apply(lambda x: [find_user_mention(user_mention) for user_mention in x['user_mentions']])

Example output with random data:

0                       [(NunkMasKKs, 🍣 SUSHIPLANERO 🍣)]
1                          [(leobilanski, Leo Bilanski)]
2                      [(romerodiario, El Profe Romero)]
3                              [(HugoYasky, Hugo Yasky)]
4                    [(marianorecalde, Mariano Recalde)]
5                     [(cyngarciaradio, Cynthia García)]