Can you use the ignoreChanges when you read your stream? The code would look something like

import dlt
from pyspark.sql.functions import col, expr
 
@dlt.view
def users():
    return (
        spark.readStream
        .format("delta")
        .option("ignoreChanges", "true")
        .table("cdc_data.users")
)
 
dlt.create_target_table("target")
 
dlt.apply_changes(
  target = "target",
  source = "users",
  keys = ["userId"],
  sequence_by = col("sequenceNum"),
  apply_as_deletes = expr("operation = 'DELETE'"),
  except_column_list = ["operation", "sequenceNum"]
)