Split a string in pandas row and insert new rows by enlarging the dataframe

Question

I have the following DataFrame:

	no	word	status
0	0	one	to_check
1	1	two	to_check
2	2	:)	emoticon
3	3	dr.	to_check
4	4	"future"	to_check
5	5	to	to_check
6	6	be	to_check

I want to iterate trough each row to find quotes at word initial and final positions and create a DataFrame like this:

	no	word	status
0	0	one	to_check
1	1	two	to_check
2	2	:)	emoticon
3	3	dr.	to_check
4	4	"	quotes
5	4	future	word
6	4	"	quotes
7	5	to	to_check
8	6	be	to_check

I can strip quotes and split the word into three pieces but I got the this DataFrame, it overwrites the last two rows:

	no	word	status
0	0	one	to_check
1	1	two	to_check
2	2	:)	emoticon
3	3	dr.	to_check
4	4	"	quotes
5	4	future	word
6	4	"	quotes

I tried df.loc[index], df.iloc[index], df.at[index] but none of them helped me to extend the number of rows in the DataFrame.

Is it possible to add new rows at specific index without overwriting last two rows?

Answer 1

In your case you can split then explode

out = df.assign(word = df.word.str.split(r'(\")')).explode('word').\
           loc[lambda x : x['word']!='']
   no    word    status
0   0     one  to_check
1   1     two  to_check
2   2      :)  emoticon
3   3     dr.  to_check
4   4       "  to_check
4   4  future  to_check
4   4       "  to_check
5   5      to  to_check
6   6      be  to_check

For change the status

out['status'] = np.where(out['word'].eq('"'), 'quotes',out['status'])

Answer 2

you try creating this function.

def InsertRow(df, idx):
    df1 = df[0:idx]
    df2 = df[idx:]

    arr_empty = np.array([idx-1, '\"', 'quotes'])

    df1.iloc[-1]=arr_empty
    df = pd.concat([df1, df2])

    return df

Answer 3

This is not very efficient however I think it is very readable, maybe you could use it if you don't care too much of efficiency.

no_lst = list()
word_lst = list()
status_lst = list()

def check_quote(w):
    if w.startswith('"') and w.endswith('"'):
        return True
    else:
        return False

for i, row in enumerate(df.itertuples()):
    word = getattr(row, "word")
    status = getattr(row, "status")
    
    if check_quote(word):
        no_lst += [i,i,i]
        stripped_w = word.strip('"')
        
        word_lst.append('"')
        word_lst.append(stripped_w)
        word_lst.append('"')
        
        status_lst += ["quotes", "word", "quotes"]
        continue
        
    no_lst.append(i)
    word_lst.append(word)
    status_lst.append(status)

new_df = pd.DataFrame({"no":no_lst,
                   "word":word_lst,
                   "status":status_lst})