如何将数据帧转换为json

Question

如何将dataframe转成json结构？

df = 

name    size    imports
AAA     111     [222,333]
BBB     111     [333,444]

结果应该是：

json = 
[
{"name":"AAA","size":111,"imports":["data.222","data.333"]},
{"name":"BBB","size":111,"imports":["data.33","data.444"]}
]

如果我使用此命令df.reset_index().to_json(orient='records')，则无法自定义data.333 之类的值（即添加data 并将值放入".."。

Answer 1

您可以使用apply 和list comprehension 将文本data. 添加到list 的每个值：

df = pd.DataFrame({'name':['AAA','BBB'],
                   'sizeB':[111,111],
                   'imports':[[222,333],[333,444]]})

df.imports = df.imports.apply(lambda x: ['data.'+ str(y) for y in x])
print (df)
                imports name  sizeB
0  [data.222, data.333]  AAA    111
1  [data.333, data.444]  BBB    111

print (df.to_json(orient='records'))
[{"imports":["data.222","data.333"],"name":"AAA","sizeB":111},
 {"imports":["data.333","data.444"],"name":"BBB","sizeB":111}]