In the use of hive Data Warehouse large data query, there is a common problem is that the query is slow, can not give users a quick data analysis query.
For decision-makers, how to get the data of user behavior analysis at the second level is a topic,
The previous approach was to import the data into hive, analyze the statistics, and then go back to other databases (such as MongoDB),
The user can make a data query based on this database.
But this has been a data processing time, (ETL data processing process) Our practice is to do every night
Data statistical analysis, and then synchronized to the database.
Now share a case of technology selection and architecture implementation of a data warehouse using Impala-kudu architecture
The technology selection and architecture realization of the analysis of divine Strategy
url:https://sensorsdata.cn/blog/technical_implementation_of_sensors_analytics/