Change word in dataframe to different word from different dataframe if they match-CodePudding

I have some trouble with my dataframe comparison. What I have are two dataframes, the first has tokenised words.

df_1:
id  sentence               some more info 
1   [I, am, happy]         bla 
2   [I, am, happier]       bla 
3   [I, am, the, saddest]  bla

and

df_2:
id word   more     most 
1  happy  happier  happiest 
2  sad    sadder   saddest

What I want to do is compare the two dataframes and if a word in df_1 matches a word anywhere in df_2 that it will be changed to df_2['word'] in the row of the corresponding word. So my output would look something like this:

df_1
id  sentence               some more info new_sentence
1   [I, am, happy]         bla       [I, am, happy]
2   [I, am, happier]       bla       [I, am, happy]
3   [I, am, the, saddest]  bla       [I, am, the, sad]

I have tried some things using .compare() and writing a function, but nothing has seemed to work so far.

Thanks for your help in advance!

CodePudding user response：

Create dictionary from second DataFrame by remove id column, reshape by DataFrame.melt and DataFrame.set_index:

d = df.drop('id', axis=1).melt('word').set_index('value')['word'].to_dict()

And then map values in dict.get with return same values if no match:

df_1['new_sentence'] = df_1['sentence'].apply(lambda x: [d.get(y, y) for y in x])

Or:

d = df.drop('id', axis=1).melt('word').set_index('value')['word'].to_dict()

df_1['new_sentence'] = [[d.get(y, y) for y in x] for x in df_1['sentence']]