在 numpy 数组中找到仅符号不同的行对

find pairs of rows in numpy array that differ only by sign

我需要在 numpy 数组中找到仅符号不同的所有行的行索引。例如,如果我有数组:

>>> A
array([[ 0,  1,  2],
       [ 3,  4,  5],
       [ 0, -1, -2],
       [ 9,  5,  6],
       [-3, -4, -5]])

我希望输出为 [(0,2),(1,4)]

我知道如何找到唯一的行,numpy.unique,所以我的直觉是将数组附加到自身的否定,即 numpy.concatenate(A,-1*A),然后找到非唯一行,但我对如何从中提取我需要的信息感到困惑。此外,数组可能非常大,因此将其附加到自身可能不是一个好主意。

通过遍历数组并检查行索引是否等于另一个行索引的否定,我得到了正确的答案,但这需要很长时间。我想要和 numpy.unique.

一样快的东西

我已经删除了 A 中的所有重复行,如果这对流程有任何影响的话。

您可以在 one-liner:

[(i, j) for i in range(len(a)) for j in range(i+1, len(a)) if all(abs(a[i]) == abs(a[j]))]

你的 a 给出:

[(0, 2), (1, 4)]

所以我们基本上是使用嵌套 for-loops 来遍历每一对 rows - ij。然后我们检查第一个 row 中的每个元素(使用 all)是否等于(==)另一个 row 中的每个元素。但是,为了引入绝对方面,我们只是在比较之前先取每个row中的abs()


哦,还有一个确切的 negation

[(i, j) for i in range(len(a)) for j in range(i+1, len(a)) if all(a[i] == -a[j])]

对于此示例给出相同的输出,但对于其他 arrays.

显然会发生变化

尝试:

A = [[0,1,2],[3,4,5],[0,-1,-2],[9,5,6],[-3,-4,-5]]

outlist = []
c = 1
while len(A) > 1:
    b = list(map(lambda x: -x, A[0]))
    A = A[1:]
    for i in range(len(A)):
        if A[i] == b:
            outlist.append((c-1, c+i))
    c += 1
print(outlist)

输出:

[(0, 2), (1, 4)]

这是一个主要基于 NumPy 的 -

def group_dup_rowids(a):
    sidx = np.lexsort(a.T)
    b = a[sidx]
    m = np.concatenate(([False], (b[1:] == b[:-1]).all(1), [False] ))
    idx = np.flatnonzero(m[1:] != m[:-1])
    C = sidx.tolist()
    return [C[i:j] for i,j in zip(idx[::2],idx[1::2]+1)]

out = group_dup_rowids(np.abs(a))

样本运行-

In [175]: a
Out[175]: 
array([[ 0,  1,  2],
       [ 3,  4,  5],
       [ 0, -1, -2],
       [ 9,  5,  6],
       [-3, -4, -5]])

In [176]: group_dup_rowids(np.abs(a))
Out[176]: [[0, 2], [1, 4]]

精确否定案例

对于寻找完全否定配对匹配的情况,我们只需要稍作修改-

def group_dup_rowids_negation(ar):
    a = np.abs(ar)
    sidx = np.lexsort(a.T)
    b = ar[sidx]
    m = np.concatenate(([False], (b[1:] == -b[:-1]).all(1), [False] ))
    idx = np.flatnonzero(m[1:] != m[:-1])
    C = sidx.tolist()
    return [(C[i:j]) for i,j in zip(idx[::2],idx[1::2]+1)]

样本运行-

In [354]: a
Out[354]: 
array([[ 0,  1,  2],
       [ 3,  4,  5],
       [ 0, -1, -2],
       [ 9,  5,  6],
       [-3, -4, -5]])

In [355]: group_dup_rowids_negation(a)
Out[355]: [[0, 2], [1, 4]]

In [356]: a[-1] = [-3,4,-5]

In [357]: group_dup_rowids_negation(a)
Out[357]: [[0, 2]]

运行时测试

其他工作方法-

# @Joe Iddon's soln
def for_for_if_listcompr(a):
    return [(i, j) for i in range(len(a)) for j in range(i+1, len(a)) 
            if all(a[i] == -a[j])]

# @dkato's soln
def find_pairs(A):
  res = []
  for r1 in range(len(A)):
    for r2 in range(r1+1, len(A)):
      if all(A[r1] == -A[r2]):
        res.append((r1, r2))
  return res

计时 -

In [492]: # Setup bigger input case
     ...: import pandas as pd
     ...: np.random.seed(0)
     ...: N = 2000 # datasize decider
     ...: a0 = np.random.randint(0,9,(N,10))
     ...: a = a0[np.random.choice(len(a0),4*N)]
     ...: a[np.random.choice(len(a),2*N, replace=0)] *= -1
     ...: a = pd.DataFrame(a).drop_duplicates().values

In [493]: %timeit for_for_if_listcompr(a)
     ...: %timeit find_pairs(a)
1 loop, best of 3: 6.1 s per loop
1 loop, best of 3: 6.05 s per loop

In [494]: %timeit group_dup_rowids_negation(a)
100 loops, best of 3: 2.05 ms per loop

进一步改进

def group_dup_rowids_negation_mod1(ar):
    a = np.abs(ar)
    sidx = np.lexsort(a.T)
    b = ar[sidx]
    dp = view1D(b)
    dn = view1D(-b)
    m = np.concatenate(([False], dp[1:] == dn[:-1], [False] ))
    return zip(sidx[m[1:]], sidx[m[:-1]])

def group_dup_rowids_negation_mod2(ar):
    a = np.abs(ar)
    sidx = lexsort_cols_posnum(a)
    b = ar[sidx]
    dp = view1D(b)
    dn = view1D(-b)
    m = np.concatenate(([False], dp[1:] == dn[:-1], [False] ))
    return zip(sidx[m[1:]], sidx[m[:-1]])

辅助函数:

#  @Divakar
def view1D(a): # a is array
    a = np.ascontiguousarray(a)
    void_dt = np.dtype((np.void, a.dtype.itemsize * a.shape[1]))
    return a.view(void_dt).ravel()

# Used to convert each row as a scalar by considering each of them as
# an indexing tuple and getting argsort indices
def lexsort_cols_posnum(ar):
    shp = ar.max(0)+1
    s = np.concatenate((np.asarray(shp[1:])[::-1].cumprod()[::-1],[1]))
    return ar.dot(s).argsort()

运行时测试(借自@Paul Panzer 的基准测试)-

In [628]: N = 50000 # datasize decider
     ...: a0 = np.random.randint(0,99,(N,3))
     ...: a = a0[np.random.choice(len(a0),4*N)]
     ...: a[np.random.choice(len(a),2*N, replace=0)] *= -1
     ...: # OP says no dups
     ...: a = np.unique(a, axis=0)
     ...: np.random.shuffle(a)

In [629]: %timeit use_unique(a) # @Paul Panzer's soln
10 loops, best of 3: 33.9 ms per loop

In [630]: %timeit group_dup_rowids_negation(a)
10 loops, best of 3: 54.1 ms per loop

In [631]: %timeit group_dup_rowids_negation_mod1(a)
10 loops, best of 3: 37.4 ms per loop

In [632]: %timeit group_dup_rowids_negation_mod2(a)
100 loops, best of 3: 17.3 ms per loop

这是 Joe Iddon 发布的功能版本。主要区别在于 if 语句:如果一对 [1, 2, 3] 和 [-1, 2, 3] 是正确的,那么我认为 Joe 的 if 语句是正确的。

def find_pairs(A):
  res = []
  for r1 in range(len(A)):
    for r2 in range(r1+1, len(A)):
      if all(A[r1] == -A[r2]):
        res.append((r1, r2))
  return res

这是一个基于 np.unique 的快速解决方案。 numpy1.13 是必需的。

import numpy as np

# Divakar's method for reference
def group_dup_rowids_negation(ar):
    a = np.abs(ar)
    sidx = np.lexsort(a.T)
    b = ar[sidx]
    m = np.concatenate(([False], (b[1:] == -b[:-1]).all(1), [False] ))
    idx = np.flatnonzero(m[1:] != m[:-1])
    C = sidx.tolist()
    return [(C[i:j]) for i,j in zip(idx[::2],idx[1::2]+1)]


def use_unique(a):
    sign = np.sign(a)
    nz = np.flatnonzero(sign)
    firstnz = np.searchsorted(nz, np.arange(0, a.size, a.shape[1]))
    a_nrm = np.where(sign.ravel()[nz[firstnz], None]==-1, -a, a)
    uniq, idx, inv, cnt = np.unique(a_nrm, True, True, True, axis=0)
    dup = np.flatnonzero(cnt==2)
    out = np.empty((len(dup), 2), dtype=int)
    out[:, 0] = idx[dup]
    idx[inv] = np.arange(len(inv))
    out[:, 1] = idx[dup]
    return out


N = 50000 # datasize decider
a0 = np.random.randint(0,99,(N,3))
a = a0[np.random.choice(len(a0),4*N)]
a[np.random.choice(len(a),2*N, replace=0)] *= -1
# OP says no dups
a = np.unique(a, axis=0)
np.random.shuffle(a)

idxd = np.array(group_dup_rowids_negation(a))
idxp = use_unique(a)

assert len(idxd) == len(idxp)
assert not np.any(np.sum(a[idxd, :], axis=1))
assert not np.any(np.sum(a[idxp, :], axis=1))
assert {frozenset(i) for i in idxd} == {frozenset(i) for i in idxp}

from timeit import timeit

gl = {'a': a}
for fun, tag in [(group_dup_rowids_negation, 'D '), (use_unique, 'pp')]:
    gl['f'] = fun
    print(tag, timeit('f(a)', number=10,  globals=gl))

示例输出:

D  0.5263204739894718
pp 0.3610327399801463

既然你说你的数组A是独一无二的,那么这个呢?

import itertools as it

In [3]: idxs_comb = list(it.combinations(range(A.shape[0]), 2))
In [4]: rows_comb = it.combinations(A, 2)
In [5]: [idxs_comb[idx] for idx, pair in enumerate(rows_comb) if np.sum(pair) == 0]
Out[6]: [(0, 2), (1, 4)]