如何避免在 Pandas 数据帧上写一个粗的 for 循环

Question

I have a DataFrame like:

import pandas as pd

begin_month = pd.Series([1, 19, 45, 32, 54])
end_month = pd.Series([19,45,32,54,99])

inventory = pd.DataFrame({"begin_month":begin_month, "end_month": end_month})

I want to make a third column, a boolean value, that says, "for each month, does the begin_month inventory == the previous month's end_month inventory level?"

I can write a foul for-loop that does this, but am wondering how I could write a vectorized action to achieve the same thing. Furthermore, the edge case is index location 0, for which there is nothing to compare its begin_month value to.

Answer 1

import pandas as pd

begin_month = pd.Series([1, 19, 145, 32, 54])
end_month = pd.Series([19,45,32,54,99])

df = pd.DataFrame({"begin_month":begin_month, "end_month": end_month})

df['parity'] = df['begin_month'] == df['end_month'].shift()
df.ix[0,'parity'] = True

print df

The key is to use .shift() so that you can compare the current row with an adjacent row. and I set df.ix[0, 'parity'] = True because it has no predecessor to compare it to.