我有一个有点复杂，或许并不复杂，但很长的问题，我将尝试提炼到基本部分，然后我可以从那里弄明白。
所以我要做的就是填补一个棒球名单。我有一个玩家列表，然后根据位置分成几个其他列表。从这一点来说，我想用每种可能的球员组合来填补名单。我写了一个基本的脚本来完成这个，但是由于我有一个大的列表，简单地遍历每个组合都不太理想。我使用了一部分玩家来测试它，它仍然需要大约一个半小时。
有些玩家有资格被列入多个位置的名单，因此，有些名字会出现在两个或多个名单中。我第一次尝试保存尝试检查每次迭代是否包含重复项，如果是，则跳过该播放器。
import os, csv, time
player_pool_csv = open('Available Player Pool.csv', 'r') player_pool = csv.reader(player_pool_csv) player_pool = list(player_pool)
roster = ['a','b','c','d','e','f','g','h']
# roster with characters is my solution to the TypeError: unhashable type: 'list'
# when I check len(roster) against len(set(roster)) in the first loop
catcher_pool = []
first_pool = []
second_pool = []
third_pool = []
short_pool = []
of_pool = []
for player in player_pool:
if 'C' in player[2]:
catcher_pool.append(player)
if '1B' in player[2]:
first_pool.append(player)
if '2B' in player[2]:
second_pool.append(player)
if '3B' in player[2]:
third_pool.append(player)
if 'SS' in player[2]:
short_pool.append(player)
if 'OF' in player[2]:
of_pool.append(player)
start = time.time()
for catcher in catcher_pool:
roster[0] = catcher[0]
if len(roster) != len(set(roster)):
continue
for first_baseman in first_pool:
roster[1] = first_baseman[0]
if len(roster) != len(set(roster)):
continue
for second_baseman in second_pool:
roster[2]= second_baseman[0]
if len(roster) != len(set(roster)):
continue
for third_baseman in third_pool:
roster[3] = third_baseman[0]
if len(roster) != len(set(roster)):
continue
for shortstop in short_pool:
roster[4] = shortstop[0]
if len(roster) != len(set(roster)):
continue
for outfielder1 in of_pool:
roster[5] = outfielder1[0]
if len(roster) != len(set(roster)):
continue
for outfielder2 in of_pool:
roster[6] = outfielder2[0]
if len(roster) != len(set(roster)):
continue
for outfielder3 in of_pool:
roster[7] = outfielder3[0]
if len(roster) != len(set(roster)):
continue
print(roster)
end = time.time()
elapsed = end - start
print(elapsed)
当我运行此代码时，它似乎在某个时刻在“for outfielder2”循环期间停止。我似乎无法确切地说明它为什么会停止它，以及如何解决它。
我意识到你可能需要更多的信息，比如它正在拉动玩家的游泳池的内容，但如果事实证明不是必要的话，我不想让这个问题超载。如果需要，我会把那些东西放进去。
任何想法我做错了什么，以及如何提高效率？感谢。
修改
好的，我把我的player_pool减去了
['Ender Inciarte', '0.283', 'OF', '3900']
['A.J. Pollock', '0.304', 'OF', '4900']
['Jamie Romak', '0.349', 'OF', '2000']
['Adam Jones', '0.258', 'OF', '3700']
['Paul Goldschmidt', '0.343', '1B', '5600']
['Chris Davis', '0.306', '1B', '4300']
['Aaron Hill', '0.245', '2B/3B', '2400']
['Jimmy Paredes', '0.276', '1B/2B', '3000']
['Jake Lamb', '0.283', '3B', '3700']
['Manny Machado', '0.315', '3B', '4400']
['Welington Castillo', '0.31', 'C', '3800']
['Caleb Joseph', '0.266', 'C', '3200']
['Xander Bogaerts', '0.318', '3B/SS', '4300']
['Eugenio Suarez', '0.294', 'SS', '3000']
当我运行所有if len(roster) != len(set(roster))注释掉的代码时，它会正确返回每个组合（每写入一个计时器需要48.3秒）。但是如果我重新放入if len(roster) != len(set(roster))，这就是输出：
['Welington Castillo', 'Paul Goldschmidt', 'Aaron Hill', 'Jake Lamb', 'Xander Bogaerts', 'Ender Inciarte', 'A.J. Pollock', 'Jamie Romak']
['Welington Castillo', 'Paul Goldschmidt', 'Aaron Hill', 'Jake Lamb', 'Xander Bogaerts', 'Ender Inciarte', 'A.J. Pollock', 'Adam Jones']
['Welington Castillo', 'Paul Goldschmidt', 'Aaron Hill', 'Jake Lamb', 'Xander Bogaerts', 'Ender Inciarte', 'Jamie Romak', 'A.J. Pollock']
['Welington Castillo', 'Paul Goldschmidt', 'Aaron Hill', 'Jake Lamb', 'Xander Bogaerts', 'Ender Inciarte', 'Jamie Romak', 'Adam Jones']
0.03500199317932129
答案 0 :(得分：2)
您可以通过已经在外部循环中选择的玩家过滤内部循环池来优化很多。套装尤其容易
你也可以使用itertools的combinations为你的外场球员避免另外三个循环
结果如下：
import itertools as it
import time
player_pool = [['Ender Inciarte', '0.283', 'OF', '3900'],
['A.J. Pollock', '0.304', 'OF', '4900'],
['Jamie Romak', '0.349', 'OF', '2000'],
['Adam Jones', '0.258', 'OF', '3700'],
['Paul Goldschmidt', '0.343', '1B', '5600'],
['Chris Davis', '0.306', '1B', '4300'],
['Aaron Hill', '0.245', '2B/3B', '2400'],
['Jimmy Paredes', '0.276', '1B/2B', '3000'],
['Jake Lamb', '0.283', '3B', '3700'],
['Manny Machado', '0.315', '3B', '4400'],
['Welington Castillo', '0.31', 'C', '3800'],
['Caleb Joseph', '0.266', 'C', '3200'],
['Xander Bogaerts', '0.318', '3B/SS', '4300'],
['Eugenio Suarez', '0.294', 'SS', '3000']]
# create a player hash for later use, I hope player names are unique
player_hash = {p[0]: p for p in player_pool}
# use sets for the pools so that we can use set difference later on
catcher_pool = set()
first_pool = set()
second_pool = set()
third_pool = set()
short_pool = set()
of_pool = set()
# fill the pools only with player names
for player, stats in player_hash.items():
if 'C' in stats[2]:
catcher_pool.add(player)
if '1B' in stats[2]:
first_pool.add(player)
if '2B' in stats[2]:
second_pool.add(player)
if '3B' in stats[2]:
third_pool.add(player)
if 'SS' in stats[2]:
short_pool.add(player)
if 'OF' in stats[2]:
of_pool.add(player)
start = time.time()
playing = set()
all_rosters = []
roster = [None]*8
# create a little generator that only yields players from a pool
# which are currently not playing
def filtered(pool):
for player in pool-playing:
playing.add(player)
yield player
playing.remove(player)
with open('rosters.txt', 'w') as f:
for catcher in filtered(catcher_pool):
roster[0] = catcher
for first_baseman in filtered(first_pool):
roster[1] = first_baseman
for second_baseman in filtered(second_pool):
roster[2] = second_baseman
for third_baseman in filtered(third_pool):
roster[3] = third_baseman
for shortstop in filtered(short_pool):
roster[4] = shortstop
for outfielders in it.combinations(of_pool-playing, 3):
roster[5:8] = outfielders
# append result to a list instead of printing
# this is a lot faster
all_rosters.append(roster[:])
# if our list is bigger than 1e6, write it to file
if len(all_rosters) > 10**6:
f.writelines(repr(roster)+'\n'
for roster
in all_rosters)
all_rosters = []
# don't forget to write the last batch which did not reach 1e6
f.writelines(repr(roster)+'\n' for roster in all_rosters)
end = time.time()
elapsed = end - start
# make sure no roster contains double names
assert all(len(set(roster)) == len(roster) == 8 for roster in all_rosters)
# make sure all rosters are unique
assert len(all_rosters) == len(set(tuple(roster) for roster in all_rosters))
print(len(all_rosters))
print(elapsed)
打印：
[a lot of rosters]
232
0.00205016136169
但结果是不有序。我认为这不是一个问题。如果要对它们进行排序，只需在生成它们之后对其进行排序。