我想使用 DLP 分析 BigQuery 中的表格。有可能的 ?怎么做 ?
答案 0 :(得分:5)
有可能。您需要定义storage_config才能使用BigQuery。
如果要将结果保存在另一个表中,请在作业配置中添加一个save_findings
操作。如果不执行任何操作,则只能通过projects.dlpJobs.get
方法访问作业的查找摘要。
按照python中的示例调用DLP来扫描BigQuery:
client_dlp = dlp_v2.DlpServiceClient.from_service_account_json(JSON_FILE_NAME)
inspect_job_data = {
'storage_config': {
'big_query_options': {
'table_reference': {
'project_id': GCP_PROJECT_ID,
'dataset_id': DATASET_ID,
'table_id': TABLE_ID
},
'rows_limit':10000,
'sample_method':'RANDOM_START',
},
},
'inspect_config': {
'info_types': [
{'name': 'ALL_BASIC'},
],
},
'actions': [
{
'save_findings': {
'output_config':{
'table':{
'project_id': GCP_PROJECT_ID,
'dataset_id': DATASET_ID,
'table_id': '{}_DLP'.format(TABLE_ID)
}
}
},
},
]
}
operation = client_dlp.create_dlp_job(parent=client_dlp.project_path(GCP_PROJECT_ID), inspect_job=inspect_job_data)
还有一个查询以分析结果:
client_bq = bigquery.Client.from_service_account_json(JSON_FILE_NAME)
# Perform a query.
QUERY = (
'WITH result AS ('
'SELECT'
' c1.info_type.name,'
' c1.likelihood,'
' content_locations.record_location.record_key.big_query_key.table_reference as bq,'
' content_locations.record_location.field_id as column '
'FROM '
' `'+ GCP_PROJECT_ID +'.'+ DATASET_ID +'.'+ TABLE_ID +'_DLP` as c1 '
'CROSS JOIN UNNEST(c1.location.content_locations) AS content_locations '
'WHERE c1.likelihood in (\'LIKELY\',\'VERY_LIKELY\'))'
'SELECT r.name as info_type, r.likelihood, r.bq.project_id, r.bq.dataset_id,'
' r.bq.table_id, r.column.name, count(*) as count FROM result r GROUP By 1,2,3,4,5,6 '
'ORDER By COUNT DESC'
)
query_job = client_bq.query(QUERY) # API request
rows = query_job.result()
for row in rows:
print('RULES: {} ({}) | COLUMN: {}.{}.{}:{} | count->{}'.format
(row.info_type, row.likelihood, row.project_id,row.dataset_id,row.table_id,row.name, row.count)
您可以找到更多详细信息here
答案 1 :(得分:0)
已经发布了有关您的用例的社区教程:dlp-to-datacatalog-tags。
在此之后,您可以在所有Big Query资源中运行DLP,并在Google数据目录中自动创建标签。
因此,您可以使用Google数据目录搜索语法来搜索敏感信息。