mdbq 0.0.2__tar.gz → 0.0.4__tar.gz
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- {mdbq-0.0.2 → mdbq-0.0.4}/PKG-INFO +1 -1
- mdbq-0.0.4/mdbq/aggregation/__init__.py +4 -0
- mdbq-0.0.2/mdbq/pipeline/database.py → mdbq-0.0.4/mdbq/aggregation/aggregation.py +38 -49
- mdbq-0.0.4/mdbq/clean/__init__.py +4 -0
- mdbq-0.0.4/mdbq/clean/data_clean.py +1321 -0
- {mdbq-0.0.2 → mdbq-0.0.4}/mdbq/company/copysh.py +1 -4
- {mdbq-0.0.2 → mdbq-0.0.4}/mdbq.egg-info/PKG-INFO +1 -1
- {mdbq-0.0.2 → mdbq-0.0.4}/mdbq.egg-info/SOURCES.txt +4 -2
- {mdbq-0.0.2 → mdbq-0.0.4}/setup.py +1 -1
- mdbq-0.0.2/mdbq/pipeline/__init__.py +0 -4
- {mdbq-0.0.2 → mdbq-0.0.4}/README.txt +0 -0
- {mdbq-0.0.2 → mdbq-0.0.4}/mdbq/__init__.py +0 -0
- {mdbq-0.0.2 → mdbq-0.0.4}/mdbq/bdup/__init__.py +0 -0
- {mdbq-0.0.2 → mdbq-0.0.4}/mdbq/bdup/bd.py +0 -0
- {mdbq-0.0.2 → mdbq-0.0.4}/mdbq/company/__init__.py +0 -0
- {mdbq-0.0.2 → mdbq-0.0.4}/mdbq/config/__init__.py +0 -0
- {mdbq-0.0.2 → mdbq-0.0.4}/mdbq/config/get_myconf.py +0 -0
- {mdbq-0.0.2 → mdbq-0.0.4}/mdbq/config/update_conf.py +0 -0
- {mdbq-0.0.2 → mdbq-0.0.4}/mdbq/log/__init__.py +0 -0
- {mdbq-0.0.2 → mdbq-0.0.4}/mdbq/log/mylogger.py +0 -0
- {mdbq-0.0.2 → mdbq-0.0.4}/mdbq/mongo/__init__.py +0 -0
- {mdbq-0.0.2 → mdbq-0.0.4}/mdbq/mongo/mongo.py +0 -0
- {mdbq-0.0.2 → mdbq-0.0.4}/mdbq/mysql/__init__.py +0 -0
- {mdbq-0.0.2 → mdbq-0.0.4}/mdbq/mysql/mysql.py +0 -0
- {mdbq-0.0.2 → mdbq-0.0.4}/mdbq/other/__init__.py +0 -0
- {mdbq-0.0.2 → mdbq-0.0.4}/mdbq/other/pov_city.py +0 -0
- {mdbq-0.0.2 → mdbq-0.0.4}/mdbq/other/ua_sj.py +0 -0
- {mdbq-0.0.2 → mdbq-0.0.4}/mdbq/other/xigua_porxy.py +0 -0
- {mdbq-0.0.2 → mdbq-0.0.4}/mdbq/pbix/__init__.py +0 -0
- {mdbq-0.0.2 → mdbq-0.0.4}/mdbq/pbix/pbix_refresh.py +0 -0
- {mdbq-0.0.2 → mdbq-0.0.4}/mdbq/pbix/refresh_all.py +0 -0
- {mdbq-0.0.2 → mdbq-0.0.4}/mdbq/spider/__init__.py +0 -0
- {mdbq-0.0.2 → mdbq-0.0.4}/mdbq.egg-info/dependency_links.txt +0 -0
- {mdbq-0.0.2 → mdbq-0.0.4}/mdbq.egg-info/top_level.txt +0 -0
- {mdbq-0.0.2 → mdbq-0.0.4}/setup.cfg +0 -0
@@ -156,9 +156,6 @@ class DataTypes:
|
|
156
156
|
class DatabaseUpdate:
|
157
157
|
def __init__(self, path):
|
158
158
|
self.path = path # 数据所在目录, 即: 下载文件夹
|
159
|
-
self.up_to_mogo: bool = False # 不设置则不上传 mongodb
|
160
|
-
self.up_to_mysql: bool = False # 不设置则不上传 mysql
|
161
|
-
self.up_to_company_mysql: bool = False # 公司台式机更新
|
162
159
|
self.datas: list = [] # 带更新进数据库的数据集合
|
163
160
|
|
164
161
|
def cleaning(self, is_move=True):
|
@@ -764,53 +761,43 @@ class DatabaseUpdate:
|
|
764
761
|
os.remove(os.path.join(root, name))
|
765
762
|
json_data.dtypes_to_file() # 写入 json 文件, 包含数据的 dtypes 信息
|
766
763
|
|
767
|
-
def upload_df(self):
|
764
|
+
def upload_df(self, service_databases=[{}]):
|
768
765
|
"""
|
769
766
|
将清洗后的 df 上传数据库
|
770
767
|
"""
|
771
|
-
|
772
|
-
|
773
|
-
|
774
|
-
database
|
775
|
-
|
776
|
-
|
777
|
-
|
778
|
-
|
779
|
-
|
780
|
-
|
781
|
-
|
782
|
-
|
783
|
-
|
784
|
-
|
785
|
-
|
786
|
-
|
787
|
-
|
788
|
-
|
789
|
-
username=username,
|
790
|
-
password=password,
|
791
|
-
host=host,
|
792
|
-
port=port,
|
793
|
-
)
|
794
|
-
if self.up_to_company_mysql: # 公司台式机更新
|
795
|
-
username, password, host, port = get_myconf.select_config_values(
|
796
|
-
target_service='company',
|
797
|
-
database='mysql',
|
798
|
-
)
|
799
|
-
cmp = mysql.MysqlUpload(
|
800
|
-
username=username,
|
801
|
-
password=password,
|
802
|
-
host=host,
|
803
|
-
port=port,
|
804
|
-
)
|
768
|
+
for service_database in service_databases:
|
769
|
+
for service_name, database in service_database.items():
|
770
|
+
# print(service_name, database)
|
771
|
+
if database == 'mongodb':
|
772
|
+
username, password, host, port = get_myconf.select_config_values(
|
773
|
+
target_service=service_name,
|
774
|
+
database=database,
|
775
|
+
)
|
776
|
+
d = mongo.UploadMongo(
|
777
|
+
username=username,
|
778
|
+
password=password,
|
779
|
+
host=host,
|
780
|
+
port=port,
|
781
|
+
drop_duplicates=False,
|
782
|
+
)
|
783
|
+
for data in self.datas:
|
784
|
+
df, db_name, collection_name = data['数据主体'], data['数据库名'], data['集合名称']
|
785
|
+
d.df_to_mongo(df=df, db_name=db_name, collection_name=collection_name)
|
805
786
|
|
806
|
-
|
807
|
-
|
808
|
-
|
809
|
-
|
810
|
-
|
811
|
-
|
812
|
-
|
813
|
-
|
787
|
+
elif database == 'mysql':
|
788
|
+
username, password, host, port = get_myconf.select_config_values(
|
789
|
+
target_service=service_name,
|
790
|
+
database=database,
|
791
|
+
)
|
792
|
+
m = mysql.MysqlUpload(
|
793
|
+
username=username,
|
794
|
+
password=password,
|
795
|
+
host=host,
|
796
|
+
port=port,
|
797
|
+
)
|
798
|
+
for data in self.datas:
|
799
|
+
df, db_name, collection_name = data['数据主体'], data['数据库名'], data['集合名称']
|
800
|
+
m.df_to_mysql(df=df, db_name=db_name, tabel_name=collection_name)
|
814
801
|
|
815
802
|
def convert_df_cols(self, df):
|
816
803
|
""""
|
@@ -967,11 +954,9 @@ class DatabaseUpdate:
|
|
967
954
|
|
968
955
|
def main():
|
969
956
|
d = DatabaseUpdate(path='/Users/xigua/Downloads')
|
970
|
-
d.up_to_mogo = True
|
971
|
-
d.up_to_mysql = True
|
972
957
|
d.new_unzip(is_move=True)
|
973
958
|
d.cleaning(is_move=False)
|
974
|
-
d.upload_df()
|
959
|
+
d.upload_df(service_databases=[{'home_lx': 'mongodb'}, {'home_lx': 'mysql'}])
|
975
960
|
# print(d.datas)
|
976
961
|
|
977
962
|
|
@@ -1046,6 +1031,10 @@ def upload():
|
|
1046
1031
|
if __name__ == '__main__':
|
1047
1032
|
username, password, host, port = get_myconf.select_config_values(target_service='aliyun', database='mongodb')
|
1048
1033
|
print(username, password, host, port)
|
1034
|
+
|
1035
|
+
d = DatabaseUpdate(path='/Users/xigua/Downloads')
|
1036
|
+
d.upload_df(service_databases=[{'home_lx': 'mongodb'},])
|
1037
|
+
|
1049
1038
|
# main()
|
1050
1039
|
# upload()
|
1051
1040
|
# path = '/Users/xigua/数据中心/原始文件2/月数据/流量来源-自助取数-月数据'
|