Python- Pandas, filter take last element in group, then first element in group

Question

Follow up question from here: drop first and last row from within each group

In pandas, how do you drop the last row in the first groupby then the first row for all subsequent entries in the group?

e.g

      X   Y
a a   0   1
  a   2   3
  c   4   5
  d   6   7
b e   8   9
  f  10  11
  g  12  13
c h  14  15
  i  16  17
d j  18  19

I want this

    X   Y
a d 6   7
b e 8   9
c h 14 15
d j 18 19

when you say "I want this" you mean you want to drop those right ? — Rayhane Mama, Jul 13 '17 at 13:12
Yeah, i want to drop all other rows. Apologies for not being more clear! — dmin, Jul 13 '17 at 13:13
and do you have your first indexes in a list or something ? or are they kind of random ? — Rayhane Mama, Jul 13 '17 at 13:19

jezrael · Accepted Answer · 2017-07-13T13:41:02.617

3

First check first value of first level by get_level_values and then groupby with apply - first group by tail and all another by head:

first = df.index.get_level_values(0)[0]
df = df.groupby(level=0, sort=False, group_keys=False)
       .apply(lambda x: x.tail(1) if x.name == first else x.head(1))
print (df)
      X   Y
a d   6   7
b e   8   9
c h  14  15
d j  18  19

edited Jul 13 '17 at 13:41

answered Jul 13 '17 at 13:36

jezrael

822,522
95
1,334
1,252

@dmin this is a much much better solution. +1 – Scott Boston Jul 13 '17 at 13:38

Python- Pandas, filter take last element in group, then first element in group

1 Answers1