When should you choose Apache Commons Collections over Guava collections?

Both libraries extend the JDK's collection APIs, but they emphasize different things, so the right choice depends on what a project already has and what specific structure it actually needs.

1. Existing dependency - if a codebase already pulls in Commons Collections (common in older Spring or enterprise Java projects), reusing it avoids adding another large library just for one feature.
2. Unique data structures - Commons Collections offers structures Guava doesn't ship directly, like a simple copy-counting Bag, BidiMap, MultiKeyMap, and array-backed CircularFifoQueue.
3. Functional composition style - Commons Collections' Predicate/Transformer/Closure model predates Java 8 and still works cleanly on older codebases that haven't adopted java.util.function or streams.

Guava tends to be favored for new projects because of its broader adoption, richer immutable collection factories (ImmutableList, ImmutableMap), built-in caching via CacheBuilder, and generally more active day-to-day maintenance - but this is a genuine trade-off rather than a settled rule, and teams reasonably land on either library depending on what's already in place and which specific feature they need.