Comment on page


Step 1 in the execution pipeline for all bulk read and some bulk write endpoints.
A pre-filter is the first step to execute in the operator pipeline for many bulk read/write requests in RecallGraph. This filter acts in two ways:
  1. 1.
    As a compulsory entity selector that determines the scope within which to search for associated entities (events, commands or historical versions), and an _id pattern that every vertex/edge in that scope must match in order for its associated entities to proceed further down the pipeline. This is determined by the path parameter in the request.
  2. 2.
    As Transaction-Time (time of record for an event) bounds which include:
    1. 1.
      An optional lower bound (inclusive) - since,
    2. 2.
      An optional upper bound (exclusive) - until
Conditions defined by the pre-filters apply to individual documents (scope) and events (time bounds). They do not apply to groups, which may be created later in the operator execution pipeline.
Pre-filters are applied within the DB query that is used to fetch intermediate results, and so it is desirable to use them to narrow down intermediate results as much as possible, before they are processed further in the rest of the pipeline.