4.1. Visualizing Execution Time of SPARQL Query Engines

In a classical analysis of the performance of query engines, plots to compare the overall execution time of the different query engines per query are presented. With the diefpy package, it is possible to generate these plots from the metrics file using the diefpy.plot_execution_time method.

import diefpy

COLORS = ["#ECC30B", "#D56062", "#84BCDA"]

# Load the result of the other metrics (execution time, etc.) from FigShare.
metrics = diefpy.load_metrics("https://ndownloader.figshare.com/files/9660316")

# Plot the execution times of all queries and query engines as a bar chart from the metrics file.
diefpy.plot_execution_time(metrics, COLORS, log_scale=True).show()

Conclusion: For most of the queries, the performance of the different approaches is comparable in terms of execution time. nLDE Not Adaptive is not able to produce results for the queries Q4.sparql and Q5.sparql. Additionally, the nLDE Selective outperforms the other approaches for query Q2.sparql while exhibiting the worst performance for query Q17.sparql.