wuhao/data_cleaning
1
0
Fork 0
Code Issues Pull Requests Releases Wiki Activity

update

Browse Source
This commit is contained in:
kf_wuhao 2021-02-24 18:07:25 +08:00
parent edc4919c8d
commit b5de055dca
2 changed files with 13 additions and 4 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

9
task/login_repair.py
View File

@ -16,7 +16,7 @@ class LoginRepair(Task):
class Model(GBaseModel): class Model(GBaseModel):
cdate: int = Field(..., title='当天0点') cdate: int = Field(..., title='当天0点')
# first_device_id: str = Field(None, title='第一次登录设备id', alias='_first_device_id') first_device_id: str = Field(None, title='第一次登录设备id', alias='_first_device_id')
manufacturer: str = Field(None, title='设备品牌', alias='_manufacturer') manufacturer: str = Field(None, title='设备品牌', alias='_manufacturer')
model: str = Field(None, title='型号', alias='_model') model: str = Field(None, title='型号', alias='_model')
os_version: str = Field(None, title='系统版本', alias='_os_version') os_version: str = Field(None, title='系统版本', alias='_os_version')
@ -48,8 +48,11 @@ class LoginRepair(Task):
# 还没有记录的 # 还没有记录的
role_set = set(role_list) - set(exists_role_list) role_set = set(role_list) - set(exists_role_list)
for role_id in role_set: for role_id in role_set:
item = self.local_db[source_coll].find_one( # logger.debug(f'add {role_id}')
{'_game_role_id': role_id, '_event_name': 'TimeSpending'}, projection) item = self.local_db['user'].find_one(
{'_game_role_id': role_id}, projection)
if not item:
continue
try: try:
item['cdate'] = int(pd.Timestamp(item['_event_time'], unit='s', tz=self.timezone) \ item['cdate'] = int(pd.Timestamp(item['_event_time'], unit='s', tz=self.timezone) \
.normalize().timestamp()) .normalize().timestamp())

8
task/summary3.py
View File

@ -260,12 +260,18 @@ class Summary3(Task):
self.local_db[f'summary_{cat}'].update_one({'_id': data['_id']}, {'$set': data}) self.local_db[f'summary_{cat}'].update_one({'_id': data['_id']}, {'$set': data})
self.remote_db[f'summary_{cat}'].update_one({'_id': data['_id']}, {'$set': data}) self.remote_db[f'summary_{cat}'].update_one({'_id': data['_id']}, {'$set': data})
# 计算累计充值 # 计算累计充值
type_map = {
'role': '_game_role_id',
'account': '_channel_uid',
'device': '_device_id',
}
for id_, doc in update_doc.items(): for id_, doc in update_doc.items():
pipeline = [ pipeline = [
{ {
'$match': { '$match': {
"cdate": {'$gte': doc['cdate'], '$lte': cdate}, "cdate": {'$gte': doc['cdate'], '$lte': cdate},
f'_game_{cat}_id': {'$in': doc[f'now_pay_{cat}_list']}, type_map[cat]: {'$in': doc[f'now_pay_{cat}_list']},
} }
}, },
{'$group': {'_id': None, 'sum_money': {'$sum': '$money'}}} {'$group': {'_id': None, 'sum_money': {'$sum': '$money'}}}