I have the following dataframe:
And I would like to convert the columns "antecedents" and "consequents" to string, removing the "frozenset({ ... })" format and thus have, for all the rows:
"VENTOLIN S.INAL200D 100MCG", instead of frozenset({ "VENTOLIN S.INAL200D 100MCG" }).
I managed to achieve the result with:
prod = []
for i in df["antecedents"]:
prod.append(str(i))
new_set = {x.replace('frozenset', ''
).replace('})', ''
).replace('({', ''
).replace("'", "") for x in prod}
Is there a more pythonic solution?
First convert values to tuples or lists and then use DataFrame.explode
:
df = pd.DataFrame({
'antecedents':[frozenset({'aaa', 'bbb'})] * 3 + [frozenset({'nbb'})] * 3,
'consequents':[frozenset({'ccc'})] * 3 + [frozenset({'nbb', 'ddd'})] * 3,
'C':[1,3,5,7,1,0],
})
#print (df)
cols = ['antecedents','consequents']
df[cols] = df[cols].applymap(lambda x: tuple(x))
print (df)
antecedents consequents C
0 (bbb, aaa) (ccc,) 1
1 (bbb, aaa) (ccc,) 3
2 (bbb, aaa) (ccc,) 5
3 (nbb,) (nbb, ddd) 7
4 (nbb,) (nbb, ddd) 1
5 (nbb,) (nbb, ddd) 0
df1 = (df.explode('antecedents')
.reset_index(drop=True)
.explode('consequents')
.reset_index(drop=True))
print (df1)
antecedents consequents C
0 bbb ccc 1
1 aaa ccc 1
2 bbb ccc 3
3 aaa ccc 3
4 bbb ccc 5
5 aaa ccc 5
6 nbb nbb 7
7 nbb ddd 7
8 nbb nbb 1
9 nbb ddd 1
10 nbb nbb 0
11 nbb ddd 0