Combine Pandas data frame column values into new column

You can use the property that summing will concatenate the string values, so you could call fillna and pass an empty str and the call sum and pass param axis=1 to sum row-wise:

In [26]:

df['Combined_ID'] = df.fillna('').sum(axis=1)
df
Out[26]:
  ID_1 ID_2 ID_3 Combined_ID
0  abc  NaN  NaN         abc
1  NaN  def  NaN         def
2  NaN  NaN  ghi         ghi
3  NaN  NaN  jkl         jkl
4  NaN  mno  NaN         mno
5  pqr  NaN  NaN         pqr

If you’re only interested in those 3 columns you can just select them:

In [39]:

df['Combined_ID'] = df[['ID_1','ID_2','ID_3']].fillna('').sum(axis=1)
df
Out[39]:
  ID_1 ID_2 ID_3 Combined_ID
0  abc  NaN  NaN         abc
1  NaN  def  NaN         def
2  NaN  NaN  ghi         ghi
3  NaN  NaN  jkl         jkl
4  NaN  mno  NaN         mno
5  pqr  NaN  NaN         pqr