|
This is actually part of the field of data warehouse modeling. Why do I mention it here Let’s first talk about some common forms of target tables. Full table, slice table, zipper table. Full scale The full scale is the same as the name, that is, all data is synchronized to the target. Try it for tables with little data changes, such as synchronization code tables. Slice table The slicing table is divided into incremental slicing and full slicing. Full slicing means placing all daily business data in the current day's partition. Incremental slicing means only placing the current day's delta in the current day's partition. Zipper list The zipper table is the most complex. There needs to be a unique key, and you need to know whether the business data has changed. After the
change, add a new entry in the target table to record the start time Austria WhatsApp Number and end time of the changed data. Some also have fields such as version and current status (zipper table). It also depends on the time granularity of synchronization. If the time granularity is finer, there may be situations where the data cannot be synchronized to the target). Why do I say it here Because data integration products need to functionally support the table creation types of these target tables. Full synchronization of the entire scale. The incremental slicing of the slicing table needs to be able to filter out daily incremental data. Does the complex logic of the zipper table need to be logic solidified (I have only seen the logic solidification of the zipper list in Powercenter.

I have also designed a wizardstyle zipper list logic solidification). during the data synchronization process. Not only can the data be moved to the target, but the data also needs to be moved to the target in a reasonable target table structure. Interactive forms of transportation During the transportation process, there are generally three forms of interaction scripting, draganddrop, and wizard. Script style As the name suggests, scripting means writing a script to synchronize data. This form is more technical and is generally not involved too much in product design. Common scripted synchronization The oldest one is Sqoop , which implements batch data migration between structured data and Hadoop.
|
|