FastRFD is the first algorithm for discovering all valid and minimal RFDs with relaxations in value equality and constraint satisfaction simultaneously.
- Java 11 or later
- Maven 3.1.0
After building the project with maven, you can get FastRFD.jar. There are 11 command options you can choose:
-f = < > | String | Default: ""| input file
-r = < > | Integer | Default: -1 | row limit ( -1 for all rows )
-d = < > | Integer | Default 10000 | pliShard ( only for mode 8 )
-i | Boolean| Default false | output diff-set to file
-o | Boolean | Default false | output the result
-g | Boolean | Default true | g1 = 0 (false)
-t = < > | Double | Default 0.01 | g1 error
-s = < > | String | Default "" | similarity threshold file
-e = < > | String | Default "" | diff-set file (if exists)
-m = < > | Integer | Default 1 | mode
-k = < > | Integer | Default 50 | topK number ( only for mode 6)
And there are modes you can choose:
mode [1] | FastRFD
mode [2] | DiffBuilder + ColEnum
mode [4] | DiffBuilder + RFDD-
mode [6] | Topk
mode [7] | ColRFD
mode [8] | rowRFD
When you want to run our code, you should choose options you need. Among these, [-f]、[-s] are necessary to input.
For example, you can run this code to find RFDs in "iris.csv" with it's given similarity threshold file "threshold/iris.txt" and g1 error 0.1, by using our algorithm FastRFD.
java -jar FastAFD.jar -f ./dataset/iris.csv -t 0.1 -s threshold/iris.txt
FastRFD are compared to other four discovery methods, Domino , Dime , pyro and FastDD. The source code of Domino can be found here. The source code of Dime can be found here. The source code of pyro can be found here. The source code of FastDD can be found here.
FastRFD is released under the Apache 2.0 license.