SDM-Genomic-Dataset [Draft]

This benchmark is created by randomly sampling data records from somatic mutation data collected in COSMIC (https://cancer.sanger.ac.uk/cosmic). SDM-Genomic-Datasets include eight different logical sources of various sizes, including 10k, 100k, 1M, and 5M rows. For every pair of sources of the same size, they differ in the percentage of the data duplicate rate, which can be either 25% or 75%, where each duplicate value is repeated 20 times. For example, a 10k logical source with 25% data duplicate rate has 75% duplicate-free records (i.e., 7500 rows), and the rest of the 25% records (i.e., 2500 rows) correspond to 125 different records which are duplicated 20 times; in total there are 7625 unique values.