An unformatted event table is created when you issue a CREATE EVENT MONITOR statement that includes the clause WRITE TO UNFORMATTED EVENT TABLE. The column definitions are useful when you want to extract data to analyze or prune a table of unneeded data.
There is no automatic purging of the event data written to an unformatted event table. You must manually purge data from the table. The column definitions for the unformatted event table are useful when you want to purge a targeted set of records. Another option is to remove all the table rows using the TRUNCATE TABLE statement.
As part of the CREATE EVENT MONITOR statement, you can specify what to name the associated unformatted event table. If not specified, the name defaults to the same name as the event monitor. The SYSCAT.EVENTTABLES catalog view lists event monitors, their associated unformatted table, and other details.
The table below describes the columns in the unformatted event table. The key column is the event_data column. The other columns represent identifiers that you can use to locate events of interest. For further attributes of table columns, issue a DESCRIBE statement.
Column name | Column data type | Column description |
---|---|---|
appl_id | VARCHAR | The identifier of the application within which the event occurred. A NULL value indicates that the application ID was not available. |
appl_name | VARCHAR | The name of the application within which the event occurred. A NULL value indicates that the application name was not available. |
event_correlation_id | BIT DATA | An optional event correlation ID. A NULL
value indicates that the event correlation ID was not available. The value is based on the event monitor type:
|
event_data | BLOB | The entire event record data for an event captured by the event monitor, stored in its original binary form. |
event_id | INTEGER | For locking event monitor records, an event identifer that is unique across the database. The ID is recycled at database activation time. Uniqueness is guaranteed by the combination of event_timestamp, event_id, member, and event_type. For UOW event monitor records, an alias of the UOW ID that is unique per connection. Uniqueness is guaranteed by the combination of event_timestamp, event_id, member, event_type and appl_id. |
event_timestamp | TIMESTAMP | The timestamp when the event was generated by the event monitor All child records will share the same timestamp as the parent record. |
event_type | VARCHAR |
The event type that occurred at the member of detection. |
member | SMALLINT | The member where the event occurred. |
partitioning_key | INTEGER | The partitioning key for the table, so that insert operations are performed locally on the database partition where the event monitor is running. |
record_seq_num | INTEGER | The sequence number of the record that is stored within the event_data column. |
record_type | INTEGER | The type of record that is stored within the event_data column. |
service_subclass_name | VARCHAR | The name of the service subclass within which the event occurred. A NULL value indicates that the service subclass name was not available. |
service_superclass_name | VARCHAR | The name of the service superclass within which the event occurred. A NULL value indicates that the service superclass name was not available. |
workload_name | VARCHAR | The name of the workload within which the event occurred. A NULL value indicates that the workload name was not available. |