Record Events

In the Event Stream Processor, streams process "record events." An event is a record, which is a composite value that associates field names with values, and an operation (for example, insert or update). Some of the fields may be designated as key fields.

• Record Types
  A record type defines the structure of the record, including the name and datatype of each field in the record.
• Record Values
  When you have a named record, you can assign it values - either static or variable.
• Key Fields
  Streams store at most one record for each unique key. That is, the values in the key field or fields must be unique within the stream.
• Record Casting
  Records are implicitly coerced depending on their context. Extra fields are dropped and missing fields are made null.
• Hidden Fields
  Records have three implicit fields, ROWID, ROWTIME and BIGROWTIME, of datatype long, date and bigdatetime respectively. Streams fill in these values automatically, and you can access them with the usual . operation.
• Operations
  Implicit within each event is an operation that is either insert, update, delete, upsert, or safedelete.