我有三个 Pandas 列，其中元素是 list。为了组合这些列表，我可以通过显式写列的名称和 + 将它们放在一起
df = pd.DataFrame({'allmz':([[1,2,3],[2,4,5],[2,5,5],[2,3,5],[1,4,5]]),'allint':([[11,31,31],[21,41,51],[41,51,51],[11,31,51],[1,51,11]]), 'allx':([[6,7,3],[2,4,5],[2,5,5],[2,9,5],[3,4,5]])})
df['new'] = df['allmz'] + df['allint'] + df['allint']
print df
allint allmz allx new
0 [11, 31, 31] [1, 2, 3] [6, 7, 3] [1, 2, 3, 11, 31, 31, 11, 31, 31]
1 [21, 41, 51] [2, 4, 5] [2, 4, 5] [2, 4, 5, 21, 41, 51, 21, 41, 51]
2 [41, 51, 51] [2, 5, 5] [2, 5, 5] [2, 5, 5, 41, 51, 51, 41, 51, 51]
3 [11, 31, 51] [2, 3, 5] [2, 9, 5] [2, 3, 5, 11, 31, 51, 11, 31, 51]
4 [1, 51, 11] [1, 4, 5] [3, 4, 5] [1, 4, 5, 1, 51, 11, 1, 51, 11]
但是，如果我有太多的列名来写每个列名，有没有办法通过循环(或不循环)列名列表来做到这一点: columns = ['allmz','allint','allx'] 代替？
请您参考如下方法:
选项 1
在列上切片并沿第一个轴调用 sum。
df['new'] = df[['allmz','allint','allx']].sum(axis=1)
df
allint allmz allx new
0 [11, 31, 31] [1, 2, 3] [6, 7, 3] [1, 2, 3, 11, 31, 31, 6, 7, 3]
1 [21, 41, 51] [2, 4, 5] [2, 4, 5] [2, 4, 5, 21, 41, 51, 2, 4, 5]
2 [41, 51, 51] [2, 5, 5] [2, 5, 5] [2, 5, 5, 41, 51, 51, 2, 5, 5]
3 [11, 31, 51] [2, 3, 5] [2, 9, 5] [2, 3, 5, 11, 31, 51, 2, 9, 5]
4 [1, 51, 11] [1, 4, 5] [3, 4, 5] [1, 4, 5, 1, 51, 11, 3, 4, 5]
选项 2
np.concatenate 的另一种选择:
v = df[['allmz','allint','allx']].values.tolist()
df['new'] = np.concatenate(v, axis=0).reshape(len(df), -1).tolist()
df
allint allmz allx new
0 [11, 31, 31] [1, 2, 3] [6, 7, 3] [1, 2, 3, 11, 31, 31, 6, 7, 3]
1 [21, 41, 51] [2, 4, 5] [2, 4, 5] [2, 4, 5, 21, 41, 51, 2, 4, 5]
2 [41, 51, 51] [2, 5, 5] [2, 5, 5] [2, 5, 5, 41, 51, 51, 2, 5, 5]
3 [11, 31, 51] [2, 3, 5] [2, 9, 5] [2, 3, 5, 11, 31, 51, 2, 9, 5]
4 [1, 51, 11] [1, 4, 5] [3, 4, 5] [1, 4, 5, 1, 51, 11, 3, 4, 5]
