General requirements

• Maven (mvn)
• a modern Java version (e.g. Java 21)
• config.json (or supply the path to your config.json as command line argument; see below for configuration parameters)

Sample dataset

• To allow easy testing, we packaged all necessary files for the Twitter dataset.
• Please import the four data CSVs (profile.csv, profile_insertiontime.csv, posts.csv, posts_insertiontime.csv) into your PostgreSQL database.
• Place the other three CSVs (rule_twitter.csv, schema_twitter.csv, derived_twitter.csv) in the configPath or link them directly in your config.json (see below).
• If no config.json is supplied and the three CSVs are placed in the same folder as the Erasure.jar, experiment 1 for the Twitter dataset is run.
• The dataset can be downloaded here.

Configuration Parameters Documentation

All configuration parameters are optional. The defaults noted in ConfigParameter.java are used if no configuration option is given.

Dataset and File Paths

• dataset: Specifies the dataset to be used.
• configPath: Path to the configuration folder.
• ruleFile: Path to the rules file, if the file name is not derivable from dataset.
• schemaFile: Path to the schema file, if the file name is not derivable from dataset.
• derivedFile: Path to the file that marks derived data attributes, if the file name is not derivable from dataset.
• All table definition in the ruleFile, schemaFile, derivedFile must contain a schema name.

Database Connection

• connectionUrl: The URL for connecting to the PostgreSQL database. Default value is "jdbc:postgresql://localhost:5432/".
• database: The name of the database to connect to. Default value is "database".
• username: The username for the database connection. Default value is "postgres".
• password: The password for the database connection. Default value is "postgres".

Processing Parameters

• numKeys: The number of keys to be processed.
• batching: Boolean flag to enable or disable the batching experiments.
• scheduling: Boolean flag to enable or disable the scheduling experiments.
• The batching and scheduling together control which experiment is executed.

• 		batching
  		true	false
  scheduling	true	experiment 5	experiment 4
  	false	experiment 3	experiment 1

• averageDependence: Boolean flag to enable or disable average dependence calculation, i.e., experiment 2.
• batchSizes: An array of integers specifying the batch sizes.
• isBatchSizeTime: Boolean flag to indicate if the batch sizes are interpreted as times.
• measureMemory: Boolean flag to enable or disable memory measurement.
• startSchedule: The start time for the schedule.
• endSchedule: The end time for the schedule.
• baseFrequency: The base frequency for scheduling.