Partitioning and Secondary Indexes

Secondary indexes allow searching by something other than the primary key (e.g., search cars by color). They don't partition cleanly.

Document-Partitioned Indexes (Local)

Each partition maintains its own secondary index, covering only the documents in that partition.

• Write: Fast. Only update the local partition.
• Read: Slow. Scatter/Gather. You must query all partitions and merge results.

Term-Partitioned Indexes (Global)

A global index covers data from all partitions. The index itself is partitioned (e.g., term "red" is on Node A, "silver" on Node B).

• Read: Fast. You only query the node containing the term "red".
• Write: Slow/Complex. A write to a document might require a distributed transaction to update indexes on different nodes.