将字典映射标签分配给pandas中的列的索引值

Question

我想使用'cluster1'字典来填充一个名为pandas series的空Cluster。

有关数据的一些背景知识：

data.StringTogeth1
print(type(data.StringTogeth1))
print(type(data.StringTogeth1[0]))

输出：

    0                      love dog cute think
    1                           dog look weird
    2                  think look like cupacak
    3    want snowman thank jerk grrr make mad
    4        hey know time babi shark dodododo
    5                               dog awesom

    Name: StringTogeth1, dtype: object


<class 'pandas.core.series.Series'>
<class 'str'>

输入：

nclusters1 = 4
clusters1 = cluster_sentences(data.StringTogeth1, nclusters1)
data['Cluster'] = pd.Series()


print(clusters1)

输出：

{1: [0, 2], 2: [1, 5], 0: [3], 3: [4]}

所以clusters1翻译成

{cluster number: [index from series,index from series], etc..}

所以数据框应如下所示：

    Id  StringTogeth1                           Cluster
0   1   love dog cute think                     1
1   2   dog look weird                          2
2   3   think look like cupacak                 1
3   4   want snowman thank jerk grrr make mad   0
4   5   hey know time babi shark dodododo       3
5   6   dog awesom                              2

Answer 1

IIUC，你可以压扁和反转clusters1，然后分配回来：

df['Clusters'] = pd.Series({v: k for k, V in clusters1.items() for v in V})
df

   Id                          StringTogeth1  Clusters
0   0                    love dog cute think         1
1   1                         dog look weird         2
2   2                think look like cupacak         1
3   3  want snowman thank jerk grrr make mad         0
4   4      hey know time babi shark dodododo         3
5   5                             dog awesom         2

哪里，

{v: k for k, V in clusters1.items() for v in V}
# {0: 1, 1: 2, 2: 1, 3: 0, 4: 3, 5: 2}

...是索引到群集标签的映射。

专业提示：无需使用空系列初始化列。