Data Flow->> Multicast

Multicast in Chinese means multicast or multicast. The natural thing about this component is that you can input a database into the number of receive components. Here's a blog post about multicast's main role and application scenario: http://www.cnblogs.com/biwork/p/3328838.html. The original text is quoted here:

First, take a portion of the data from the same data source and put it directly into table A, and part of the data is placed directly into table B. I have encountered a file of more than 370 columns, these 370 columns of files can be divided into columns of 7, 8 tables. My first practice is to load the data of this file into a large table, and then from the large table to extract the different columns into the small table, which is equivalent to the formation of two times load, inefficient, and then use multicast can be used to directly spread the 370 column width of the file into different tables.

Second, extracting data from data source A to B,b will be Truncate each time, but you need to back up every time you extract data from a, you can use multicast at this time. Each time a data is extracted from a, the data is directed to B's backup table by multicast.

The third, similar to the second one, is that there is no backup table, but some Audit information data that is loaded is required to be preserved. For example, when extracting data from Source to Staging, it is also necessary to record the number of rows extracted and the name of the table and column that is used to indicate the maximum timestamp in the Staging data. In this case, the next time the data is loaded into Staging, only the new source data will be loaded, that is, after the last batch of the maximum timestamp.

Data Flow->> Multicast