Hadoop to build a data warehouse practice reading notes [1]

The number of positions of benefits:

  • Multiple data sources into a single data storage, so you can use a single data query engine data display.
  • Alleviate resource competition on the large transaction processing database queries arising from the implementation of children generated.
  • Maintain historical data.
  • By integrating a plurality of data source system, so that there is no uniform central view angle in the entire enterprise.
  • And description provided by encoding the same, to reduce or correct data issues, data quality.
  • Consistently organize information.
  • It provides a single common data model for all data, without concern for the data source.
  • The reconstructed data, make data more meaningful to business users.
  • To complex analytical queries deliver excellent query performance without compromising operational systems.
  • Type query development decision easier.

Use of personal experience: the ability to multi-table join queries (cross-server mysql) in the hive, query large amounts of data faster. You can do data union, de-duplication and except.

Guess you like

Origin www.cnblogs.com/astride/p/11164048.html