CODEWITHSUNDEEP
pythonpandas
MARK
MARK

Reputation: 2362

Pandas: Find index of the row with second highest value

I am trying to get the index of the row with the second highest value after doing groupby but I am not getting the right result

df = pd.DataFrame({'Sp':['a','b','c','d','e','f'], 'Mt':['s1', 's1', 's2','s2','s2','s3'], 'Value':[1,2,3,4,5,6], 'count':[3,2,5,10,10,6]})

Doing this

df.iloc[df.groupby(['Mt'])['Value'].apply(lambda x: (x!=max(x)).idxmax())]

is returning

    Mt  Sp  Value   count
0   s1  a   1   3
2   s2  c   3   5
5   s3  f   6   6

For group s2 , index 3 of the original dataframe should be returned.

Upvotes: 4

Views: 5538

Answers (2)

Andy Hayden
Andy Hayden

Reputation: 375395

Since 'Value' is already sorted you can use nth:

In [11]: g = df.groupby("Mt", as_index=False)

In [12]: g.nth(-2)
Out[12]:
   Mt Sp  Value  count
0  s1  a      1      3
3  s2  d      4     10

Otherwise I'd first sort by Value, df = df.sort_values("Value").

If you want the last (if there are fewer than two in a given group), you could grab that too 

In [21]: g = df.groupby("Mt")

In [22]: res = g.nth(-1)

In [23]: res.update(g.nth(-2))

In [24]: res
Out[24]:
   Sp  Value  count
Mt
s1  a      1      3
s2  d      4     10
s3  f      6      6

A related function is tail (to get the last two elements):

In [31]: g.tail(2)
Out[31]:
   Mt Sp  Value  count
0  s1  a      1      3
1  s1  b      2      2
3  s2  d      4     10
4  s2  e      5     10
5  s3  f      6      6

Upvotes: 4

MARK
MARK

Reputation: 2362

OK I got the answer except for one thing. This code seems to work

df.iloc[df.groupby(['Mt'])['Value'].apply(lambda x: (x!=max(x)).order(ascending=False).head(1).index[0])]

The only thing I dont understand right now that even with a group of one row only that row is being returned. I was thinking that may be x!=max(x) check would exclude that row.

Upvotes: 0

Related Questions