CODEWITHSUNDEEP
pythonpandas
Relam
Relam

Reputation: 75

Row-wise lookup from other data frame

I have two data frames that I would like to combine based on certain conditions. This is the first data frame, each line represents one obversation (thus IDs occure multiple times):

df1

  ID  Count  Publication
0  A     10         1990
1  B     15         1990
2  A     17         1990
3  B     19         1991
4  A     13         1991

This is the second data frame. Here, each ID is shown only once but over time (here 1990 to 1993).

df2

  ID  1990  1991  1992  1993
0  A   1.1   1.2   1.3   1.4
1  B   2.3   2.4   2.4   2.6
2  C   3.4   3.5   3.6   3.7
3  D   4.5   4.6   4.7   4.8

My goal is to add a results column to df1, in which I multiply the value from the df1["Count"] column with the respective value (ID-Year pair) from df2, e.g. first line: "ID" A in "1990" is 1.1 multiplied with "Count" 10 = 11.

results

  ID  Count  Publication  Results
0  A     10         1990     11.0
1  B     15         1990     34.5
2  A     17         1990     18.7
3  B     19         1991     45.6
4  A     13         1991     15.6

So far I have tried multiple options using pandas .apply() function but it did not work. I have also tried to .merge() the columns from df2 to df1 based on IDs but I still fail to make the calculation afterwards (I was hoping this simplies the problem).

Question: Is there an easy an efficient way to go throug df1 row by row and "pick" the corresponding values from df2 for the calculation?

Upvotes: 3

Views: 177

Answers (2)

BENY
BENY

Reputation: 323226

Using lookup

df2.set_index('ID').lookup(df1.ID,df1.Publication.astype(str))
Out[189]: array([1.1, 2.3, 1.1, 2.4, 1.2])

df1['Results']=df2.set_index('ID').lookup(df1.ID,df1.Publication.astype(str))*(df1.Count)
df1
Out[194]: 
  ID  Count  Publication  Results
0  A     10         1990     11.0
1  B     15         1990     34.5
2  A     17         1990     18.7
3  B     19         1991     45.6
4  A     13         1991     15.6

Upvotes: 3

Zuma
Zuma

Reputation: 846

I don't really know how efficient it is, but you can do that:

df1 = df1.set_index(['ID', 'Publication'])
df2 = df2.set_index('ID').stack()
df2.index.rename(['ID', 'Publication'], inplace=True)
df1['df2_value'] = df2
df1['result'] = df1['Count'] * df1['df2_value']

Upvotes: 1

Related Questions