Datasets

Datasets

A dataset is a collection of testcases for a problem. Each problem can have multiple datasets, and exactly one of them is flagged as the live dataset — a submission is graded against the live dataset at the time of submission. Time limit, memory limit, and scoring are configured separately per dataset, so you can prepare a new set of testcases or limits alongside the one in use and switch over when ready.

Scoring modes

There are two scoring modes:

• SUM — the submission's total score is the sum of the individual testcases' scores (each contributing according to its weight).
• GROUP MIN — testcases are bundled into groups, and each group scores by the minimum result among its testcases (IOI-style subtasks: failing one case in a group fails the whole group). By default every testcase is in its own group; see Configuring Testcases for how to assign groups and weights.

Related pages

• Configuring Testcases — per-testcase properties (num, group, weight, …) and the parameter-string formats.
• Checker and Auxiliary Files — custom checkers and other per-problem files.