CODEWITHSUNDEEP
pythonpandas
Nicholas Aldershof
Nicholas Aldershof

Reputation: 265

Duplicating Values based on Series input

I'm trying to write a custom function that essentially extends replace functionality (with some domain specific concerns) Essentially I have a series that looks like:

0   1
1   2
2   4
3   4
4   4

And return values from another function (Which I can't modify unfortunately cause it's coming from another source) that looks like either a series like:

1   A
2   B
4   C

OR a dataframe that looks like this

    cola colb colc
1   A    T    F
2   B    F    F
4   C    T    T

I'm trying to return output that either looks like

0   A
1   B
2   C
3   C
4   C

or

    cola colb colc
0   A    T    F
1   B    F    F
2   C    T    T
3   C    T    T
4   C    T    T

depending on what type is returned by the function. I'm able to write a script that performs this iteratively, but I feel like there must be a more efficient, more pandas specific way of performing this operation, so before I generate a hideous nested monstrosity I figured I'd check if there's a well-supported way!

Upvotes: 1

Views: 48

Answers (2)

EdChum
EdChum

Reputation: 393943

I don't know how your data really looks like so you may need to modify my code slightly but the following works using map:

In [32]:
s.map(s1[1])

Out[32]:
0
0    A
1    B
2    C
3    C
4    C
Name: 1, dtype: object

for the second one you can perform a left merge but you have to construct a dataframe from your series:

In [41]:
pd.DataFrame(s).merge(df, left_on=[1], right_index=True, how='outer')

Out[41]:
   1 cola colb colc
0                  
0  1    A    T    F
1  2    B    F    F
2  4    C    T    T
3  4    C    T    T
4  4    C    T    T

For the above you can set the index after the merge

EDIT

Actually for your second problem it's easier to use reindex and pass the series values:

In [42]:
df.reindex(s)

Out[42]:
  cola colb colc
1               
1    A    T    F
2    B    F    F
4    C    T    T
4    C    T    T
4    C    T    T

Upvotes: 4

A.P.
A.P.

Reputation: 1149

You can use pandas.merge to do that.

s = pd.Series([1,2,4,4,4])
df = pd.DataFrame({'colA':[1,2,4], 'colB':['A','B','C']})
pd.merge(pd.DataFrame(s, columns=['series']), df, left_on='series', right_on='colA')

Output:

       series  colA colB
0       1     1    A
1       2     2    B
2       4     4    C
3       4     4    C
4       4     4    C

Upvotes: 0

Related Questions