前言
整体评价
好绝望的牛客周赛,彻底暴露了CF菜菜的本质,F题没思路,G题用置换环骗了50%, 这大概是唯一的亮点了。
A. 小红的字符串生成
思路: 枚举
a,b两字符在相等情况下比较特殊
a, b = input().split()
if a == b:print (2)print (a)print (a + a)
else:print (4)print (a)print (b)print (a + b)print (b + a)
B. 小红的非排列构造
思路: 脑筋急转弯
如果合法,就是0
如果不合法,就把第1项改为n+1(排列外的数)
n = int(input())
arr = list(map(int, input().split()))si = set()
for v in arr:if 1 <= v <= n:si.add(v)
if len(si) != n:print (0)
else:print (1)print (1, n + 1)
C. 小红的数字拆解
思路: 贪心
从高位到低位贪心即可
s = input()arr = []i = 0
while i < len(s):j = iwhile j < len(s):p = ord(s[j]) - ord('0')if p % 2 == 0:breakj += 1arr.append(s[i:j+1])i = j + 1arr.sort(key=lambda x: (len(x), x))print (*arr, sep='\n')
D. 小红的陡峭值
思路: 构造
这题是限制性很强的题, 绝对差值总和为1
对于不满足条件数组,可以立马返回 -1.
难点在于
如何构造合法数组 如何构造合法数组 如何构造合法数组
假定0元素(左右两侧都存在非0值),和左侧元素保持一致(反证法使得该假设一定成立)
那就剩下一侧有非0值的0值,如何讨论呢?
- 如果绝对差值总和为1
- 那就和左侧/右侧的那个非0值,保持一致
- 如果绝对差值总和为0
- 那就选一边构造一个比左侧/右侧刚好大1的数
n = int(input())
arr = list(map(int, input().split()))if arr == [0 for _ in range(n)]:if len(arr) == 1:print (-1)else:print (*([1] + [2] * (n - 1)), sep=' ')
else:# 计算当前的差值综合brr = [v for v in arr if v > 0]diff = 0for i in range(len(brr) - 1):diff += abs(brr[i] - brr[i+1])if diff > 1:print (-1)else:first, last = -1, -1for i in range(n):if arr[i] != 0:if first == -1:first = ilast = i # 填充中间的值pre = arr[first]for i in range(first + 1, last):if arr[i] == 0:arr[i] = preelse:pre = arr[i]# 填充两端的值if diff == 1:# 需要两端保持绝对值差值为0for i in range(first):arr[i] = arr[first]for i in range(last + 1, n):arr[i] = arr[last]print (*arr, sep=' ')else:# 需要两端构建出一个绝对值差值1if first > 0:for i in range(first):arr[i] = arr[first] + 1for i in range(last + 1, n):arr[i] = arr[last]print (*arr, sep=' ')elif last + 1 < n:for i in range(last + 1, n):arr[i] = arr[last] + 1print (*arr, sep=' ')else:print (-1)
E. 小红的树形 dp
思路: BFS + 验证
属于诈骗题,因为题目一直再诱导你往树形DP上去靠
其实从bfs出发,进行交叉染色
然后对边上两点进行验证,如果存在同色,即不合法
n = int(input())
s = input()
ss = list(s)g = [[] for _ in range(n)]for _ in range(n - 1):u, v = list(map(int, input().split()))u -= 1v -= 1g[u].append(v)g[v].append(u)from collections import dequedeq = deque()
for i in range(n):if ss[i] != '?':deq.append((i, ss[i]))# 需要补充这种特殊情况
if len(deq) == 0:ss[0] = 'd'deq.append((0, ss[0]))# BFS流程
while len(deq) > 0:idx, ch = deq.popleft()for v in g[idx]:if ss[v] == '?':ss[v] = 'd' if ch == 'p' else 'p'deq.append((v, ss[v]))# 验证逻辑
ok = True
for i in range(n):ch = ss[i]if ch == '?':ok = Falsebreakfor v in g[i]:if ch == 'd' and ss[v] != 'p':ok = Falsebreakif ch == 'p' and ss[v] != 'd':ok = Falsebreakif not ok:print (-1)
else:print (''.join(ss))
F. 小红的矩阵构造
思路: 异或特性
贴一下皮神的代码
代码即是说服力
n, m, x = map(int, input().split())res = [[0] * m for _ in range(n)]if x % 4 == 0:res[0][0] = res[0][1] = res[1][0] = res[1][1] = x // 4for ls in res:print(*ls)
elif x == 2:print(-1)
else:res[2][0] = res[2][1] = 1res[1][0] = res[1][2] = 1res[0][1] = res[0][2] = 1for i in [0, -1]:for j in [0, -1]:res[i][j] += (x - 6) // 4for ls in res:print(*ls)
G. 小红的元素交换
思路: 置换环 + 找规律
假如这题没有交换颜色的限制,那这题就是裸的置换环
但是实际上,这题的核心框架依旧是
置换环 置换环 置换环
具体情况需要分类讨论
-
同一置换环(a个元素)中存在两种颜色, 则交换一定为a-1
-
同一置换环(a个元素)中只存在1种颜色, 则需要借助外部的非同色,这样为a-1+2=a+1
但是这样做,只能对大致50%+
为啥呢?
因为对于纯色R的置换环,和纯色W的置换环,其实可以互相借调,因此这一组合,可以减少2次不必要的交换。
因此在原先的基础上,需要优化减掉
m i n ( 纯色 R 的群数,纯色 W 的群数 ) ∗ 2 min(纯色R的群数,纯色W的群数) * 2 min(纯色R的群数,纯色W的群数)∗2
n = int(input())
arr = list(map(int, input().split()))
s = input()from collections import Counterif [i + 1 for i in range(n)] != arr and len(Counter(s)) == 1:print(-1)
else:res = 0white, red = 0, 0for i in range(n):if arr[i] == i + 1:continueelse:# 置换环核心逻辑state = 0tmp = 0while arr[i] != i + 1:p1, p2 = i, arr[i] - 1arr[p1], arr[p2] = arr[p2], arr[p1]tmp += 1state |= 2 if s[p1] == 'R' else 1state |= 2 if s[p2] == 'R' else 1# 分类讨论置换环的纯色情况if state == 3:res += tmpelse:# 纯色,需要借调外部力量res += tmp + 2if state == 1:white += 1else:red += 1print(res - min(white, red) * 2)