CODEWITHSUNDEEP
python-3.xpandas
César Barboza
César Barboza

Reputation: 187

How to obtain the percent change from the first and last value of a DataFrame in Pandas

I would like to know how to obtain the pct_change from either column balance, exports and imports, from the first and last value of the following DataFrame? Also it would be nice if you could illustrate how to get the pct_change from two specific dates

balance       date  exports  imports
0   -45053 2008-01-01   421443   466496
1   -33453 2009-01-01   399649   433102
2   -41168 2010-01-01   445748   486916
3   -25171 2011-01-01   498862   524033
4   -33364 2012-01-01   501055   534419
5   -35367 2013-01-01   519913   555280
6   -36831 2014-01-01   518925   555756
7   -32370 2015-01-01   517161   549531
8   -43013 2016-01-01   547473   590486

Upvotes: 0

Views: 2553

Answers (2)

silgon
silgon

Reputation: 7211

First, you have to set your index as date if you want to use directly the dates. Then everything goes easily.

import pandas as pd
data = [[-45053, "2008-01-01",   421443,   466496],
        [-33453, "2009-01-01",   399649,   433102],
        [-41168, "2010-01-01",   445748,   486916],
        [-25171, "2011-01-01",   498862,   524033],
        [-33364, "2012-01-01",   501055,   534419],
        [-35367, "2013-01-01",   519913,   555280],
        [-36831, "2014-01-01",   518925,   555756],
        [-32370, "2015-01-01",   517161,   549531],
        [-43013, "2016-01-01",   547473,   590486]]

columns = ["balance","date","exports","imports"]
df=pd.DataFrame(data,columns=columns).set_index("date")
print(df.loc["2009-01-01"]/df.loc["2008-01-01"]-1)
# result
# balance   -0.257475
# exports   -0.051713
# imports   -0.071585
# dtype: float64

Upvotes: 2

Zero
Zero

Reputation: 76937

IIUIC, use iloc for first .iloc[0] and last .iloc[-1] to get pct_change 100*(last/first-1)

In [244]: cols = ['balance', 'exports', 'imports']

In [245]: 100*(df[cols].iloc[-1]/df[cols].iloc[0]-1)
Out[245]:
balance    -4.528000
exports    29.904400
imports    26.579006
dtype: float64

Upvotes: 4

Related Questions