pyodps

from odps.df import DataFrame
dfusers = DataFrame(o.get_table('odps_tc_257100_f673506e024.zhaopin_round2_user'))
df.dtypes
df.head(5)
df[df.sepalwidth > 3]['name', 'sepalwidth'].head(5)

print('获取方法表格')

print(users.schema)

print(users['experience'][1:10])
迭代

for jdStr in train_jd_data['job_description'].execute():

o.get_table('modified_zhaopin_round2_jd'

for jdStr in o.read_table('modified_zhaopin_round2_jd'):

with table.open_reader(partition=partition, **kw) as reader:

for jdStr in train_jd_data['job_description'][1]:

>>> dual.creation_time
>>> datetime.datetime(2014, 6, 6, 13, 28, 24)
>>> dual.is_virtual_view
>>> False
>>> dual.size
>>> 448
>>> dual.schema.columns

%sql select * from pyodps_iris limit 5

优秀
https://yq.aliyun.com/articles/300307?spm=a2c4e.11153940.0.0.2c623566uS5Kcg

要学的那么多，特征工程不会，spark不会


train_jd_data= DataFrame(o.get_table('modified_zhaopin_round2_jd')).to_pandas()

train_jd_data= o.get_table('modified_zhaopin_round2_jd').to_df().to_pandas()

train_jd_data= DataFrame(o.get_table('modified_zhaopin_round2_jd'))--novalue

users.exclude('zip_code', 'age').head(5)
	users.select(users.exclude('zip_code', 'sex'), sex_bool=users.sex == 'M').head(5)
value_count(
	df = users.groupby('occupation').agg(count=users['occupation'].count())
	df.sort(df['count'], ascending=False)[:10]
users[users.age.between(20, 25)].count() #年龄在20到25岁之间的人有多少个
	 users.groupby(users.sex).agg(count=users.count())
保存新表
	movies = DataFrame(o.get_table('pyodps_ml_100k_movies'))
	ratings = DataFrame(o.get_table('pyodps_ml_100k_ratings'))
	o.delete_table('pyodps_ml_100k_lens', if_exists=True)
	lens = movies.join(ratings).join(users).persist('pyodps_ml_100k_lens')
	
现在我们把年龄分成从0到80岁，分成8个年龄段。
	labels = ['0-9', '10-19', '20-29', '30-39', '40-49', '50-59', '60-69', '70-79']
	cut_lens = lens[lens, lens.age.cut(range(0, 81, 10), right=False, labels=labels).rename('年龄分组')]
	取分组和年龄唯一的前10条查看。
	cut_lens['年龄分组', 'age'].distinct()[:10]
	 cut_lens.groupby('年龄分组').agg(cut_lens.rating.count().rename('评分总数'), cut_lens.rating.mean().rename('评分均值'))
#2创建Dataframe
Collection(DataFrame) ，Sequence，Scalar。 这三个对象分别表示表结构（或者二维结构）、列（一维结构）、标量

>>> iris = DataFrame(o.get_table('pyodps_iris'))
>>> iris2 = o.get_table('pyodps_iris').to_df()  # 使用表的to_df方法
>>> df2 = DataFrame(df, unknown_as_string=True, as_type={'null_col2': 'float'})
>>> df2.dtypes
>>> #2.1，Sequence，Scalar。
>>> iris.sepallength.head(5) #获取列 getattr(df, ‘column_name’) df[column_name]
>>> #运算
>>> (iris.sepallength * 10).log().head(5)
>>> fields = [iris.sepallength,
>>>            (iris.sepallength / 2).rename('sepallength除以2'),
>>>            (iris.sepallength ** 2).rename('sepallength的平方')]