我可以根據(jù)在上一個(gè)數(shù)據(jù)流步驟中處理的數(shù)據(jù)將數(shù)據(jù)插入到不同的 bigQuery 數(shù)據(jù)集嗎？我正在創(chuàng)建一個(gè)數(shù)據(jù)流管道，它從 PubSub 訂閱中讀取數(shù)據(jù)并寫入大查詢表。其定義如下：def run(argv=None, save_main_session=True):    options: PipelineOptions = PipelineOptions(        project='project-id',        runner='DataflowRunner',        region='region',        streaming=True,        setup_file='dataflow/setup.py',        autoscaling_algorithm='THROUGHPUT_BASED',        job_name='telemetry-processing'    )    with beam.Pipeline(options=options) as p:        status = (                p                 | 'Get Status PubSub' >> beam.io.ReadFromPubSub(            subscription='projects/project-id/subscriptions/subscription-id',            with_attributes=True))        status_records = (status| 'Proto to Dict' >> beam.Map(lambda x: convert_proto_to_dict(x, nozzle_status_proto.NozzleStatus)) )        status_records | 'Write status to BQ' >> beam.io.WriteToBigQuery('project- id:dataset-id.table-id')         bytes_status = (status | 'Get Bytes Result' >> beam.ParDo(GetBytes()))         bytes_status | 'Write to BQ BackUp' >> beam.io.WriteToBigQuery(        'project-id:dataset-id.backup-table-id')對(duì)于給定的輸入和輸出，它完全按照預(yù)期工作。我想要的是，關(guān)于 PubSubMessage 中的特定屬性，定義我的消息應(yīng)該發(fā)送到哪個(gè)數(shù)據(jù)集。所以我需要改變的部分是：status_records | 'Write status to BQ' >> beam.io.WriteToBigQuery('project-id:dataset-id.table-id')我已經(jīng)嘗試提取所需的數(shù)據(jù)并像這樣使用它：status_records | 'Write status to BQ' >> beam.io.WriteToBigQuery('project-id:{data-from-previous-step}.table-id')但我們無(wú)法直接從 PCollection 獲取數(shù)據(jù)。我嘗試像這篇文章中那樣覆蓋 WriteToBigQuery（How can I write to Big Query using a runtime valueprovider in Apache Beam?），但我沒(méi)有收到錯(cuò)誤，也沒(méi)有插入任何內(nèi)容。我不知道如何實(shí)現(xiàn)這一點(diǎn)。你知道我應(yīng)該從哪里開始做這件事嗎？我是否必須為 n 個(gè)數(shù)據(jù)集創(chuàng)建 n 個(gè)管道？