[Java][Docs] Document the use of the batchSize argument in Dataset ScanOptions

Several ScanOptions methods take a batchSize argument as shown: 

`public ScanOptions(long batchSize) {`
`    this(batchSize, Optional.empty());`
``}

Since the scanner reads one ArrowRecordBatch per load invocation, setting the parameter to a size larger than the RecordBatch has no effect. It only works when it's smaller than the number of rows in the RecordBatch, (i.e., the number or records read is equal to min(batchSize, recordBatch rowCount), potentially leading to some confusion. 

**Reporter**: [Larry White](https://issues.apache.org/jira/browse/ARROW-17346) / @lwhite1

_{**Note**: *This issue was originally created as [ARROW-17346](https://issues.apache.org/jira/browse/ARROW-17346). Please see the [migration documentation](https://github.com/apache/arrow/issues/14542) for further details.*}

Commencez par les méthodes de Dataset ScanOptions qui acceptent batchSize et retracez la façon dont le scanner charge un ArrowRecordBatch par invocation. Documentez que les valeurs supérieures au nombre de lignes de RecordBatch n’ont aucun effet supplémentaire, tandis que les valeurs inférieures limitent les enregistrements à min(batchSize, recordBatch rowCount) ; le travail est terminé lorsque la documentation de ScanOptions rend ce comportement clair.

Rédigé par le modèle d'indexation à partir du texte de l'issue.