Tuesday, 28 September 2021

Replicate vs Broadcast

BROADCAST COMBINES IN AN ARBITRARY ORDER ALL RECORDS IT RECEIVES INTO A SINGLE FLOW AND WRITES A COPY OF THAT FLOW TO EACH OF ITS OUTPUT FLOW PARTITIONS. 

USE BROADCAST TO INCREASE DATA PARALLELISM WHEN YOU HAVE CONNECTED A SINGLE FAN-OUT FLOW TO THE OUT PORT, OR TO INCREASE COMPONENT PARALLELISM WHEN YOU HAVE CONNECTED MULTIPLE STRAIGHT FLOWS TO THE OUT PORT.

REPLICATE ARBITRARILY COMBINES ALL RECORDS IT RECEIVES INTO A SINGLE FLOW AND WRITES A COPY OF THAT FLOW TO EACH OF ITS OUTPUT FLOWS. USE REPLICATE TO SUPPORT COMPONENT PARALLELISM — SUCH AS WHEN YOU WANT TO PERFORM MORE THAN ONE OPERATION ON A FLOW OF RECORDS COMING FROM AN ACTIVE COMPONENT.

REPLICATE VERSUS BROADCAST-
1) REPLICATE AND BROADCAST ARE SIMILAR COMPONENTS, SO IT CAN BE DIFFICULT TO KNOW WHICH ONE TO USE IN A PARTICULAR GRAPH.

2) BROADCAST IS USED TO INCREASE DATA PARALLELISM BY FEEDING RECORDS TO FAN-OUT OR ALL-TO-ALL FLOWS.

3) REPLICATE IS GENERALLY USED TO INCREASE COMPONENT PARALLELISM, EMITTING MULTIPLE STRAIGHT FLOWS TO SEPARATE PIPELINES.

SPECIFICALLY, THE DIFFERENCE BETWEEN THEM LIES IN HOW THEIR FLOWS ARE SET UP AND HOW THEIR LAYOUTS ARE PROPAGATED IN THE GDE.

4) REPLICATE ALLOWS MULTIPLE OUTPUTS FOR A GIVEN LAYOUT AND PROPAGATES THE LAYOUT FROM THE INPUT TO THE OUTPUT.

5) BROADCAST IS A PARTITIONING COMPONENT THAT DEFINES THE TRANSITION FROM ONE LAYOUT TO ANOTHER.

No comments:

Post a Comment

how to create dml dynamically in Ab-initio

 $[ begin let int i = 0; let string(int) complete_list = "emp_nm,billable,age,designation,location"; let string(int) file_content ...