[data] Sample schemas in unification

[iamjustinhsu]

Why are these changes needed?

unification of schemas is slow. Provide a way to limit the number of schemas being sampled. When sample_size=1, unification is skipped and we just return the first non empty schema.

Related issue number

Checks

• I've signed off every commit(by using the -s flag, i.e., git commit -s) in this PR.
• I've run scripts/format.sh to lint the changes in this PR.
• I've included any doc changes needed for https://docs.ray.io/en/master/.
  • I've added any new APIs to the API Reference. For example, if I added a
    method in Tune, I've added it in doc/source/tune/api/ under the
    corresponding .rst file.
• I've made sure the tests are passing. Note that there might be a few flaky tests, see the recent failures at https://flakey-tests.ray.io/
• Testing Strategy
  • Unit tests
  • Release tests
  • This PR is not tested :(

[gemini-code-assist]

Code Review

This pull request introduces a sample_size parameter to schema unification functions to improve performance by limiting the number of schemas processed. The changes are applied across various parts of the codebase, using sample_size=1 for performance-sensitive intermediate steps and sample_size=None (for full unification) where a complete schema is necessary.

I've found a critical bug in the implementation that could lead to a TypeError when sample_size is None, which is a common case in the new code. I've also identified an opportunity to refactor duplicated code to improve maintainability. Please see my detailed comments.

[iamjustinhsu]

not needed in favor of #55926 and #55854