假设：在使用策略标记更新每个列之前和之后保持列名相同

步骤：
1.使用get_table()提取表并迭代其模式（SchemaField列表）
1.创建新的SchemaField对象，通过引用SchemaField的名称/描述（来自步骤#1）添加参数值，policy_tags除外（类型= PolicyTagList示例：google.cloud.bigquery.SchemaField(name=column.name, field_type=column.field_type, policy_tags=google.cloud.bigquery.PolicyTagList(["projects/some-project/locations/us/taxonomies/123454321/policyTags/180000"]))
1.将这些新的SchemaFields对象追加到列表中，并使用此新列表重新分配表的模式
1.使用update_table更新架构，例如：
google.cloud.bigquery.Client(project = BIGQUERY_PROJECT).update_table(table, ['schema'])

使用@Asad Siddiqui步骤，创建以下代码。
下面将介绍如何使用GCP中的云函数自动更新模式。
工艺流程：上传一个csv文件到GCS bucket中--〉用下面的代码触发云函数&应用csv文件中提到的表的模式。

• • CSV模板格式：**

import pandas as pd
from google.cloud import bigquery
bigquery_client = bigquery.Client()

def pii_auto_apply_policies(event, context):

    # READING INPUT CSV FILE TEMPLATE DATA:
    bucket_name = event['bucket']
    object_name = event['name']
    print(object_name, bucket_name)

    path = "gs://{}/{}".format(bucket_name, object_name)
    data = pd.read_csv(path, encoding='utf-8')
    print(data.head())

    for row in range(0, len(data.table_id)):

        table = bigquery_client.get_table(data.table_id[row])  # Make an API request.
        original_schema = table.schema

        # Creating new schema list to create a new one.
        new_schema = list()

        # Looping through originial schema to create a new schema list:
        for schema_row in original_schema:

            # 1st if - appends all other column schemas other than the ones required.
            if schema_row.name != data.field_name_in_table[row] or schema_row.field_type != data.data_type_in_table[row]:
                new_schema.append(schema_row)
                
            # 2nd if - picks only the required column and appends with policy tag resource id.
            if schema_row.name == data.field_name_in_table[row] and schema_row.field_type == data.data_type_in_table[row]:
                new_schema.append(bigquery.SchemaField(name=schema_row.name,
                                    field_type=schema_row.field_type,
                                    policy_tags=bigquery.PolicyTagList([data.policy_tag_resource[row]])))

        print("Schema length for table: {}, before: {}, after: {}".format(data.table_id[row], 
                                                                        len(original_schema), len(new_schema)))

        # Applying the schema to table
        table.schema = new_schema
        table = bigquery_client.update_table(table, ["schema"])
        print("Table updated:{}".format(data.table_id[row]))

    return "OK"

局限性：这不适用于"记录"类型的列（其中有嵌套字段）。我将更新第二个版本，以防万一，如果我找到了解决方案。