问题描述
我在Maxcompute使用的Mars进行数据分析 df[‘point’] = df[‘point’].map(lambda x: int(x[1:]))报错如下：
Traceback (most recent call last):
File “”, line 383, in
main_ktfh(“19:30-20:30”)
File “”, line 362, in main_ktfh
sea_df = fun_season_change(s_point, e_point)
File “”, line 327, in fun_season_change
a_df = fun_melt(fdf)
File “”, line 254, in fun_melt
df[‘point’] = df[‘point’].map(lambda x: int(x[1:]))
File “/opt/taobao/tbdpapp/pyodps/mars/0.6/mars/dataframe/base/map.py”, line 253, in series_map
return op(series, dtype=dtype)
File “/opt/taobao/tbdpapp/pyodps/mars/0.6/mars/utils.py”, line 456, in _inner
return func(*args, **kwargs)
File “/opt/taobao/tbdpapp/pyodps/mars/0.6/mars/dataframe/base/map.py”, line 104, in __call__
raise ValueError(‘cannot infer dtype, ‘
ValueError: cannot infer dtype, it needs to be specified manually for `map`
期望结果
Mars中如何实现Pandas中map、apply方法，是否能提供相关文档