Conditionally combine columns in pandas data frame

Question

I've got a DataFrame which looks something like this:

   x1  x2 
0   4   1 
1   0   2 
2   5   1 
3   0   3 
4   4   2 

Now I want to create another column which takes average of columns x1 and x2 , or returns 0 if x1 is 0 :

   x1  x2  ave
0   4   1  2.5 
1   0   2  0
2   5   1  3 
3   0   3  0
4   4   2  3

Neither this

data['ave'] = (data['x1'] + data['x2'])/2 if data['x1'] > 0 else 0

nor this

data['ave'] = (data['x1'] != 0)*(data['x1'] + data['x2'])/2

works for obvious reasons (series can't be used in these operations).

I do know that this is easy to accomplish using a loop, but is there a shorthand pythonic way of doing it?

Proper python data is below:

data = pd.DataFrame({'x1': (4,0,5,0,4), 'x2': (1,2,1,3,2)})

Answer 1

You're very close. Both of your approaches should work with only a tweak or two. Method #1:

>>> df = pd.DataFrame({'x1': (4,0,5,0,4), 'x2': (1,2,1,3,2)})
>>> df["ave"] = (df["x1"] != 0) * (df["x1"] + df["x2"])/2.
>>> df
   x1  x2  ave
0   4   1  2.5
1   0   2  0.0
2   5   1  3.0
3   0   3  0.0
4   4   2  3.0

If you leave off the . in 2. and your columns are integers, you might not get the results you expect due to integer division, but Series can be used without any problems.

Method #2:

df["ave"] = df.apply(lambda r: (r["x1"] + r["x2"])/2. if r["x1"] > 0 else 0, axis=1)

Pass a function to apply and specify axis=1 .

Method #3a, 3b:

df["ave"] = df.mean(axis=1) * (df["x1"] != 0)

or

df["ave"] = df[["x1", "x2"]].mean(axis=1)
df["ave"][df["x1"] == 0] = 0

Et cetera.