我有以下数据框:
load = pd.DataFrame({'A':list('abcdef'),
'B':[4,5,4,5,5,4],
'C':[7,8,9,4,2,0],
'D':[1,3,5,4,2,0],
'E':[5,3,6,9,2,4],
'F':list('aaabbb')})
如果条件满足我需要,我需要检查 b>c 其中 c=d 和 c,d!=0把 True else False。
输出
False
False
False
True
True
False
请您参考如下方法:
我认为需要通过 & 为 bitwise AND 链接 bool 掩码:
m = load.B.gt(load.C) & load.C.eq(load.D) & load[['C','D']].ne(0).any(axis=1)
#alternative
m = load.B.gt(load.C) & load.C.eq(load.D) & load['C'].ne(0) & load['D'].ne(0)
print (m)
0 False
1 False
2 False
3 True
4 True
5 False
dtype: bool
如果想要 numpy.where 得到相同的输出:
load['new'] = np.where(m, True, False)
print (load)
A B C D E F new
0 a 4 7 1 5 a False
1 b 5 8 3 3 a False
2 c 4 9 5 6 a False
3 d 5 4 4 9 b True
4 e 5 2 2 2 b True
5 f 4 0 0 4 b False
如果需要按条件设置 2 个值,则使用它,例如:
load['new'] = np.where(m, 10, 20)
print (load)
A B C D E F new
0 a 4 7 1 5 a 20
1 b 5 8 3 3 a 20
2 c 4 9 5 6 a 20
3 d 5 4 4 9 b 10
4 e 5 2 2 2 b 10
5 f 4 0 0 4 b 20
性能:
load = pd.concat([load] * 1000, ignore_index=True)
In [106]: %timeit load['print'] = load.apply(lambda x:(x.B>x.C)&(x.C==x.D)&(x.C!=0)&(x.D!=0),axis=1)
408 ms ± 13.3 ms per loop (mean ± std. dev. of 7 runs, 1 loop each)
In [107]: %timeit load['print'] = load.B.gt(load.C) & load.C.eq(load.D) & load['C'].ne(0) & load['D'].ne(0)
1.64 ms ± 135 µs per loop (mean ± std. dev. of 7 runs, 1000 loops each)
