Sentence selection and summary generation are two main steps to generate informative and readable summaries.However,most previous works treat them as two separated subtasks.In this paper,we propose a novel extractive-and-abstractive hybrid framework for single document summarization task by jointly learning to select sentence and rewrite summary.It first selects sentences by an extractive decoder and then generate summary according to each selected sentence by an abstractive decoder.Moreover,we apply the BERT pre-trained model as document encoder,sharing the context representations to both decoders.Experiments on the CNN/DailyMail dataset show that the proposed framework outperforms both state-of-the-art extractive and abstractive models.