使用zip形成一个新的数据框，但收到错误

Question

我有一个名为playoff_teams的numpy_array：

playoff_teams = np.sort(playoff_seeds['team'])
playoff_teams[:7]

array([1115, 1124, 1139, 1140, 1143, 1155, 1165], dtype=int64)

我有一个名为reg的数据框：

       season daynum    wteam   wscore  lteam   lscore  wloc    numot
108122  2010    7       1143     75      1293    70       H     0
108123  2010    7       1314     88      1198    72       H     0
108124  2010    7       1326     100     1108    60       H     0
108125  2010    7       1393     75      1107    43       H     0
108126  2010    9       1143     95      1178    61       H     0

然后我循环遍历团队并执行以下操作：

for teams in  playoff_teams:
    games = reg[(reg['wteam'] == teams) | (reg['lteam']== teams)]
    last_six = sum(games.tail(6)['wteam'] == teams)
    zipped = zip(team, last_six)

我收到错误

TypeError: zip argument #1 must support iteration

我需要以下列格式形成一个新的数据框：

col_1   col_2
team_1   last_six
team_2   last_six
team_3   last_six

我该怎么做？

Answer 1

sum()返回一个数字，而不是你可以迭代的东西，而zip()需要迭代，所以我认为你的问题就在那里。

last_six = sum(games.tail(6)['wteam'] == teams)  # Number
zipped = zip(team, last_six)  # Error because last_six is not iterable

您可以将结果存储在列表中（也可能是dict），例如：

new_data = []
for teams in  playoff_teams:
    games = reg[(reg['wteam'] == teams) | (reg['lteam']== teams)]
    last_six = sum(games.tail(6)['wteam'] == teams)
    new_data.append((teams, last_six))

然后使用DataFrame.from_items或DataFrame.from_dict构建您的数据框（如果您选择了dict而不是列表）。