Lifecycle Callbacks

Hooks let an object run code at each step of its own insert, update, delete, and select. A PikaBean already implements them, so override only the ones you want. The classic use is timestamps:

public class Post extends PikaBean {
    String title;
    LocalDateTime createdAt;
    LocalDateTime updatedAt;

    @Override
    public boolean beforeInsert() {
        createdAt = LocalDateTime.now();   // stamp the row on the way in
        return true;
    }

    @Override
    public boolean beforeUpdate() {
        updatedAt = LocalDateTime.now();   // a field set here is in the UPDATE
        return true;
    }
}

(If you are not using PikaBean, implement PikaRecordLifecycle directly. It is the same interface.)

The hooks

Every method has a default, so you implement only what you need.

public interface PikaRecordLifecycle {
    // pre-operation: return false to abort
    default boolean validate() { return true; }
    default boolean beforeInsert() { return true; }
    default boolean beforeUpdate() { return true; }
    default boolean beforeDelete() { return true; }

    // post-operation
    default void afterInsert() {}
    default void afterSelect() {}
    default void afterUpdate() {}
    default void afterDelete() {}
}

Order

• Insert: validate then beforeInsert then the SQL then afterInsert. Pika reads the fields after beforeInsert, so a field set in beforeInsert is in the INSERT. The generated primary key is set when afterInsert runs.
• Bulk insert (insertAll): validate then beforeInsert for each item, then one SQL statement. If a hook returns false, Pika skips that item. Pika does not call afterInsert, and does not set the generated ids.
• Update: validate then beforeUpdate then the SQL then afterUpdate. Pika reads the fields after beforeUpdate, so a field set in beforeUpdate is in the UPDATE. For a PikaBean, Pika then removes the columns that did not change.
• Delete: beforeDelete then the SQL then afterDelete.
• Select: the SQL then afterSelect, which runs once fields are mapped (this is where PikaBean snapshots state for dirty tracking).

Aborting

Return false from any pre-operation hook and Pika silently skips the operation: no SQL runs, and insert / update / delete report no rows affected. Nothing is thrown.

To actually stop the request or roll back a transaction, throw instead:

@Override
public boolean beforeDelete() {
    if (locked) throw new IllegalStateException("cannot delete a locked post");
    return true;
}

For plain field checks, use Validation & Forms. Reach for hooks when you need timestamps, cascades, or logic tied to one specific step.