What should a Java toMap collector do when two records share a key?

A duplicate key is a data-model decision, not just a collector exception to suppress. Decide whether duplicates represent an error, multiple values or a mergeable quantity. For quantities that should be summed, use a merge function such as Collectors.toMap(Row::code, Row::quantity, Integer::sum), with overflow handling if totals can grow. For multiple independent records, groupingBy preserves the group instead. Reject duplicates explicitly when each identifier must be unique; silently keeping the last value can destroy evidence. Verify two equal keys with different values, an all-unique list and the required output order. Hash-map output does not promise input order. A merge rule that depends on arrival order may produce different answers in a parallel pipeline.