Bucketing strategy
Bucketing strategy is one aspect of how the Oracle Unity stores data for a data object. It can help join performance.
In bucketing, the system splits data by the bucketing attribute, usually the ID attribute. It then spreads the data object's data across several files, determined by the number of buckets.
Managing the bucketing strategy for a data object
Important: The bucketing strategy for a data object can't be changed after it has been published. When creating a data object, make sure you are satisfied with the bucketing strategy before publishing the changes.
Number of data object records/Number of buckets
When creating a data object, you will need to estimate the number of records the data object will have, up to a maximum of two billion records. Depending on the number of records you select, the number of buckets for the data object will be automatically adjusted under Number of buckets. If you want to update the number of buckets, use the slider for Expected number of records (in millions) to adjust the number of buckets.
Bucketing attribute
The default bucketing attribute is the ID attribute. The goal for an ideal bucketing design is to spread the load equally across all the buckets, so that all of the individual buckets are equally written to and read from. However, depending on the write and read patterns, ID may not be the best attribute to use. If most of the reads and writes are happening in a single (or a few) buckets, then the load has not been spread equally. To update the bucketing attribute, you will need to first create the data object, make sure not to publish changes, and then follow the steps for Editing data objects.