要在Apache Beam中使用BigQuery的Upsert（更新或插入）功能，可以使用BigQuery的Java客户端库和Beam的ParDo转换。下面是一个示例代码：

import org.apache.beam.sdk.Pipeline;
import org.apache.beam.sdk.coders.AvroCoder;
import org.apache.beam.sdk.io.gcp.bigquery.BigQueryIO;
import org.apache.beam.sdk.transforms.DoFn;
import org.apache.beam.sdk.transforms.ParDo;
import org.apache.beam.sdk.values.PCollection;
import org.apache.beam.sdk.values.Row;
import org.apache.beam.sdk.values.TypeDescriptor;

public class UpsertExample {
  public static void main(String[] args) {
    // 创建一个Pipeline
    Pipeline pipeline = Pipeline.create();

    // 读取BigQuery表中的数据
    PCollection input = pipeline.apply(
        BigQueryIO.readTableRows().from("project-id:dataset.table"));

    // 将数据转换为UpsertRow对象
    PCollection upsertRows = input.apply(
        ParDo.of(new ConvertToUpsertRowFn())
            .withOutputType(TypeDescriptor.of(UpsertRow.class)));

    // 将UpsertRow对象写入BigQuery表中
    upsertRows.apply(
        BigQueryIO.writeTableRows()
            .to("project-id:dataset.table")
            .withFormatFunction(row -> row.toTableRow())
            .withWriteDisposition(BigQueryIO.Write.WriteDisposition.WRITE_APPEND)
            .withCreateDisposition(BigQueryIO.Write.CreateDisposition.CREATE_NEVER));

    // 运行Pipeline
    pipeline.run();
  }

  // 自定义DoFn用于将Row对象转换为UpsertRow对象
  static class ConvertToUpsertRowFn extends DoFn {
    @ProcessElement
    public void processElement(ProcessContext c) {
      Row row = c.element();
      UpsertRow upsertRow = new UpsertRow(row.getString("id"), row.getString("name"));
      c.output(upsertRow);
    }
  }

  // UpsertRow对象包含要插入或更新的数据
  static class UpsertRow {
    public String id;
    public String name;

    public UpsertRow(String id, String name) {
      this.id = id;
      this.name = name;
    }

    public TableRow toTableRow() {
      return new TableRow()
          .set("id", id)
          .set("name", name);
    }
  }
}

在这个示例中，我们首先读取了BigQuery表中的数据，并将其转换为UpsertRow对象。然后，我们使用BigQueryIO将UpsertRow对象写入BigQuery表中。在ConvertToUpsertRowFn中，我们使用自定义的DoFn将Row对象转换为UpsertRow对象。最后，我们将UpsertRow对象转换为TableRow对象，并将其写入BigQuery表中。