multimapping

Multimapping, in general terms, refers to a relation in which a single key or query is associated with multiple values or locations. It describes a one-to-many mapping where the value side is a collection rather than a single item. The term is used across fields to contrast with one-to-one mappings and to highlight ambiguity or multiplicity in the associations.

In data structures and programming, a multimap is a container that maps keys to multiple values. Each key can have several associated elements, stored as a list, vector, set, or other collection. This is different from a standard map, where each key is linked to exactly one value. Common implementations include C++'s std::multimap, which allows multiple entries with the same key, and higher-level libraries such as Java's Guava Multimap or Python patterns like a dict of lists. Typical operations include inserting a (key, value) pair, retrieving all values for a given key, and removing one or all values associated with a key. Some implementations offer range views to access all entries sharing a key efficiently.

In databases, multimapping corresponds to one-to-many relationships. A single parent key relates to multiple child records stored in a separate table with a foreign key. SQL queries join the related tables to fetch all associated values. Indexing on the foreign key improves retrieval performance. In practice, relational schemas use primary/foreign keys and often surrogate keys to model multimaps.

In genomics and bioinformatics, multimapping describes sequencing reads that align to multiple genomic locations with similar scores. This ambiguity affects downstream analyses; strategies include distributing read counts across locations, selecting a primary alignment, or reporting all possible alignments with mapping quality scores.

Key considerations include memory usage, performance implications, and the desired semantics for insertion, retrieval, and deletion.