I tried drop method of pandas but I didn't use it. How do I remove a specific row in pandas with Python?
e.g.: My specific row is => Name: Bertug Grade: A Age: 15
Add a comment
|
4 Answers
df = pd.DataFrame([['Jhon',15,'A'],['Anna',19,'B'],['Paul',25,'D']])
df. columns = ['Name','Age','Grade']
df
Out[472]:
Name Age Grade
0 Jhon 15 A
1 Anna 19 B
2 Paul 25 D
You can get the index of your row:
i = df[((df.Name == 'jhon') &( df.Age == 15) & (df.Grade == 'A'))].index
and then drop it:
df.drop(i)
Out[474]:
Name Age Grade
1 Anna 19 B
2 Paul 25 D
As @jezrael pointed our, you can also just negate all three:
df[((df.Name != 'jhon') &( df.Age != 15) & (df.Grade != 'A'))]
Out[477]:
Name Age Grade
1 Anna 19 B
2 Paul 25 D
2 Comments
cottontail
There's mistake in the negation. The negation of
A and B is ~A or ~B. I added an answer here that shows the correct negation.
Leonardo Ferreira
Beautiful to see that the De Morgan's Law is alive and well.
you can just use :
df.drop([a,b,c])
where a,b,c are the list of indexes or row numbers.
to delete only one particular row use
df.drop(i)
where i is the index or the row number.
2 Comments
Orhan Solak
You should assign to the new df, or just add inplace=True in drop command section. Without these, you could not use dropped version.
Convexity
Yes.
print(df.drop(i)) works but does not affect df. df = df.drop(i) or the like does affect.Dropping rows is the inverse operation of filtering rows you want to keep, so a negating a boolean mask that filters for the rows to drop is enough.
df[~(df['Name'].eq('Bertug') & df['Grade'].eq('A') & df['Age'].eq(15))]
which is, by de Morgan's laws, equivalent to
df[df['Name'].ne('Bertug') | df['Grade'].ne('A') | df['Age'].ne(15))]
Another approach is to simply query the rows you want to keep.
df1 = df.query("not (Name == 'Bertug' and Grade == 'A' and Age == 15)")
# or
df1 = df.query("Name != 'Bertug' or Grade != 'A' or Age != 15")
If the index of the row(s) to be dropped are given, then query (or drop) can be used too. For example, the following drops the second row.
idx = [1]
df1 = df.query("index != @idx")
# equivalently,
df1 = df.drop(idx)
Comments
you can drop just by writing this where df is your DataFrame and in index write the index of the row you want to drop df.drop(index=0, inplace=True)
1 Comment
binford
Thank you for your contribution. However, your answer does not provide anything substantial new. If you create an answer for a question that already has a high-rated answer, please think twice if your answer really adds something substantial new.