Question

我有一个pandas数据帧：

df

id  Description
1   POS Transfer A&W MONTREAL QC
2   MKLI QC Montreal DOLLARAMA
3   PC - PAYMENT FROM - *****11*22

我想从描述中获取地址。我做了什么：

provinces=["qc","on"]

b=[]
for index, row in df.iterrows():
    c=0
    for i in provinces:
        if i in row["Description"].lower().split():
            a=[row["id"],row["description"]]
            c=1
            break
    if(c==1):        
        b.append(a)


b 

[[0, 'POS Transfer A&W MONTREAL QC'],
 [1, ' MKLI QC Montreal DOLLARAMA ']]

这里我正在捕获数组中的信息。我可以直接在另外一个pandas列中捕获它吗

在这里，我捕获了有地址的所有描述。为了将此发送到google api，我只想选择该省左侧和右侧的1个单词。那就是：

POS Transfer A&W MONTREAL QC

我想抓拍：

A&W MONTREAL QC

在

的情况下

MKLI QC Montreal DOLLARAMA

我想抓拍

QC Montreal DOLLARAMA

我该怎么做？

Answer 1

我认为需要在province之前和之后选择1个单词：

provinces=["qc","on"]

def f(x):
    #convert to lowercast splitted list for find
    x1 = x.lower().split()
    x2 = x.split()
    #get indices of search word with one word before and after
    #min and max is for avoid indexing errors 
    out = [' '.join(x2[max(0, i-1): min(len(x), i+2)]) for i, y in enumerate(x1) 
                                                       if y in provinces]
    #return first matched value if exist
    return out[0] if len(out) > 0 else ''


df['new'] = df['Description'].apply(f)
print (df)
   id                     Description               new
0   1    POS Transfer A&W MONTREAL QC       MONTREAL QC
1   2      MKLI QC Montreal DOLLARAMA  MKLI QC Montreal
2   3  PC - PAYMENT FROM - *****11*22

在pandas数据框中查找列的经度和纬度

1 个答案: