删除具有重复索引的行（Pandas DataFrame和TimeSeries）

I’m reading some automated weather data from the web. The observations occur every 5 minutes and are compiled into monthly files for each weather station. Once I’m done parsing a file, the DataFrame looks something like this:

                      Sta  Precip1hr  Precip5min  Temp  DewPnt  WindSpd  WindDir  AtmPress
Date                                                                                      
2001-01-01 00:00:00  KPDX          0           0     4       3        0        0     30.31
2001-01-01 00:05:00  KPDX          0           0     4       3        0        0     30.30
2001-01-01 00:10:00  KPDX          0           0     4       3        4       80     30.30
2001-01-01 00:15:00  KPDX          0           0     3       2        5       90     30.30
2001-01-01 00:20:00  KPDX          0           0     3       2       10      110     30.28

The problem I’m having is that sometimes a scientist goes back and corrects observations — not by editing the erroneous rows, but by appending a duplicate row to the end of a file. Simple example of such a case is illustrated below:

import pandas 
import datetime
startdate = datetime.datetime(2001, 1, 1, 0, 0)
enddate = datetime.datetime(2001, 1, 1, 5, 0)
index = pandas.DatetimeIndex(start=startdate, end=enddate, freq='H')
data1 = {'A' : range(6), 'B' : range(6)}
data2 = {'A' : [20, -30, 40], 'B' : [-50, 60, -70]}
df1 = pandas.DataFrame(data=data1, index=index)
df2 = pandas.DataFrame(data=data2, index=index[:3])
df3 = df2.append(df1)
df3
                       A   B
2001-01-01 00:00:00   20 -50
2001-01-01 01:00:00  -30  60
2001-01-01 02:00:00   40 -70
2001-01-01 03:00:00    3   3
2001-01-01 04:00:00    4   4
2001-01-01 05:00:00    5   5
2001-01-01 00:00:00    0   0
2001-01-01 01:00:00    1   1
2001-01-01 02:00:00    2   2

And so I need df3 to evenutally become:

                       A   B
2001-01-01 00:00:00    0   0
2001-01-01 01:00:00    1   1
2001-01-01 02:00:00    2   2
2001-01-01 03:00:00    3   3
2001-01-01 04:00:00    4   4
2001-01-01 05:00:00    5   5

I thought that adding a column of row numbers (df3['rownum'] = range(df3.shape[0])) would help me select out the bottom-most row for any value of the DatetimeIndex, but I am stuck on figuring out the group_by or pivot (or ???) statements to make that work.

I would suggest using the duplicated method on the Pandas Index itself:

df3 = df3[~df3.index.duplicated(keep='first')]

While all the other methods work, the currently accepted answer is by far the least performant for the provided example. Furthermore, while the groupby method is only slightly less performant, I find the duplicated method to be more readable.

Using the sample data provided:

>>> %timeit df3.reset_index().drop_duplicates(subset='index', keep='first').set_index('index')
1000 loops, best of 3: 1.54 ms per loop

>>> %timeit df3.groupby(df3.index).first()
1000 loops, best of 3: 580 µs per loop

>>> %timeit df3[~df3.index.duplicated(keep='first')]
1000 loops, best of 3: 307 µs per loop

Note that you can keep the last element by changing the keep argument to 'last'.

It should also be noted that this method works with MultiIndex as well (using df1 as specified in Paul’s example):

>>> %timeit df1.groupby(level=df1.index.names).last()
1000 loops, best of 3: 771 µs per loop

>>> %timeit df1[~df1.index.duplicated(keep='last')]
1000 loops, best of 3: 365 µs per loop

This adds the index as a dataframe column, drops duplicates on that, then removes the new column:

df = df.reset_index().drop_duplicates(subset='index', keep='last').set_index('index').sort_index()

Note that the use of .sort_index() above at the end is as needed and is optional.

Oh my. This is actually so simple!

grouped = df3.groupby(level=0)
df4 = grouped.last()
df4
                      A   B  rownum

2001-01-01 00:00:00   0   0       6
2001-01-01 01:00:00   1   1       7
2001-01-01 02:00:00   2   2       8
2001-01-01 03:00:00   3   3       3
2001-01-01 04:00:00   4   4       4
2001-01-01 05:00:00   5   5       5

Follow up edit 2013-10-29 In the case where I have a fairly complex MultiIndex, I think I prefer the groupby approach. Here’s simple example for posterity:

import numpy as np
import pandas

# fake index
idx = pandas.MultiIndex.from_tuples([('a', letter) for letter in list('abcde')])

# random data + naming the index levels
df1 = pandas.DataFrame(np.random.normal(size=(5,2)), index=idx, columns=['colA', 'colB'])
df1.index.names = ['iA', 'iB']

# artificially append some duplicate data
df1 = df1.append(df1.select(lambda idx: idx[1] in ['c', 'e']))
df1
#           colA      colB
#iA iB                    
#a  a  -1.297535  0.691787
#   b  -1.688411  0.404430
#   c   0.275806 -0.078871
#   d  -0.509815 -0.220326
#   e  -0.066680  0.607233
#   c   0.275806 -0.078871  # <--- dup 1
#   e  -0.066680  0.607233  # <--- dup 2

and here’s the important part

# group the data, using df1.index.names tells pandas to look at the entire index
groups = df1.groupby(level=df1.index.names)  
groups.last() # or .first()
#           colA      colB
#iA iB                    
#a  a  -1.297535  0.691787
#   b  -1.688411  0.404430
#   c   0.275806 -0.078871
#   d  -0.509815 -0.220326
#   e  -0.066680  0.607233

Unfortunately, I don’t think Pandas allows one to drop dups off the indices. I would suggest the following:

df3 = df3.reset_index() # makes date column part of your data
df3.columns = ['timestamp','A','B','rownum'] # set names
df3 = df3.drop_duplicates('timestamp',take_last=True).set_index('timestamp') #done!

If anyone like me likes chainable data manipulation using the pandas dot notation (like piping), then the following may be useful:

df3 = df3.query('~index.duplicated()')

This enables chaining statements like this:

df3.assign(C=2).query('~index.duplicated()').mean()

Remove duplicates (Keeping First)

idx = np.unique( df.index.values, return_index = True )[1]
df = df.iloc[idx]

Remove duplicates (Keeping Last)

df = df[::-1]
df = df.iloc[ np.unique( df.index.values, return_index = True )[1] ]

Tests: 10k loops using OP’s data

numpy method - 3.03 seconds
df.loc[~df.index.duplicated(keep='first')] - 4.43 seconds
df.groupby(df.index).first() - 21 seconds
reset_index() method - 29 seconds