<< Click to Display Table of Contents >> Current:  Data Mart > MPP > ETL > Pull Data from Data Source  Accelerate Join Result

When two or more data sets are joined and incremental imported into the mart, check the "Accelerate Join Result" option to transfer the join calculation from the client node to each map node, reduce the memory pressure of the client node and improve the operation efficiency.

When the composed data set and self-service data set meet the conditions, the "Accelerate Join Result" option can be checked. If the conditions are not met, it will be grayed out,, not optional.

 

         

                                                 【Figure 1: Join result is accelerated to be selectable】

 

Meet the conditions for Join results accelerate into the market:

1. The environment needs to be greater than one Map node. This function can be selected for a single Map node or local market, but an error is reported.

2. The combination data set needs to be calculated for the join, and the last operation node before the self-service data set needs the result is "join". Otherwise, the acceleration of the "Accelerate Join Result" is not optional.

 

                 

                                                 【Figure 2】

 

3. When two or more SQL data sets from the same database are joined, ensure that the SQL will not be pushed down to the database for calculation, otherwise the "join result acceleration" setting is not optional.

 

                 

                                                 【Figure 3】

 

4. Only single-level joins are supported, and multi-level joins are not supported to accelerate the market entry. For example, (A1 join A2) join (B1 join B2) [Note: Both sides of the join are subqueries, which are multi-layer] Accelerated graying of join results is not optional.

 

         

                                                 【Figure 4: This figure is a multi-layer join】

 

When "Accelerate Join Result" is checked, only the "append" function is available.

 

                 

 

parameter:join.grid.transmit.timeout

The time limit for distributing and transmitting data when executing a task, the default is 1200000 milliseconds, an error will be reported when the timeout expires, and the task will stop. It can be configured in bi.properties.