Skip to content

Commit

Permalink
implement create_or_update_data_profile_scan method
Browse files Browse the repository at this point in the history
  • Loading branch information
syou6162 committed Nov 3, 2024
1 parent 8c49667 commit 1a13aac
Showing 1 changed file with 162 additions and 0 deletions.
162 changes: 162 additions & 0 deletions dbt/adapters/bigquery/impl.py
Original file line number Diff line number Diff line change
@@ -1,5 +1,6 @@
from dataclasses import dataclass
from datetime import datetime
import hashlib
import json
import threading
from multiprocessing.context import SpawnContext
Expand Down Expand Up @@ -53,7 +54,9 @@
import google.oauth2
import google.cloud.bigquery
from google.cloud.bigquery import AccessEntry, SchemaField, Table as BigQueryTable
from google.cloud import dataplex_v1
import google.cloud.exceptions
from google.protobuf import field_mask_pb2
import pytz

from dbt.adapters.bigquery import BigQueryColumn, BigQueryConnectionManager
Expand Down Expand Up @@ -96,6 +99,26 @@ def render(self):
return f"{self.project}.{self.dataset}"


@dataclass
class DataProfileScanSetting():
location: str
scan_id: Optional[str]

project_id: str
dataset_id: str
table_id: str

sampling_percent: Optional[float]
row_filter: Optional[str]
cron: Optional[str]

def parent(self):
return f"projects/{self.project_id}/locations/{self.location}"

def data_scan_name(self):
return f"{self.parent()}/dataScans/{self.scan_id}"


def _stub_relation(*args, **kwargs):
return BigQueryRelation.create(
database="", schema="", identifier="", quote_policy={}, type=BigQueryRelation.Table
Expand Down Expand Up @@ -999,3 +1022,142 @@ def validate_sql(self, sql: str) -> AdapterResponse:
:param str sql: The sql to validate
"""
return self.connections.dry_run(sql)

# If the label `dataplex-dp-published-*` is not assigned, we cannot view the results of the Data Profile Scan from BigQuery
def _update_labels_with_data_profile_scan_labels(
self,
project_id: str,
dataset_id: str,
table_id: str,
location: str,
scan_id: str,
):
table = self.connections.get_bq_table(project_id, dataset_id, table_id)
original_labels = table.labels
profile_scan_labels = {
"dataplex-dp-published-scan": scan_id,
"dataplex-dp-published-project": project_id,
"dataplex-dp-published-location": location,
}
table.labels = {**original_labels, **profile_scan_labels}
self.connections.get_thread_connection().handle.update_table(table, ["labels"])

# scan_id must be unique within the project and no longer than 63 characters,
# so generate an id that meets the constraints
def _generate_unique_scan_id(self, dataset_id: str, table_id: str) -> str:
md5 = hashlib.md5(f"{dataset_id}_{table_id}".encode("utf-8")).hexdigest()
return f"dbt-{table_id.replace('_', '-')}-{md5}"[:63]

def _create_or_update_data_profile_scan(
self,
client: dataplex_v1.DataScanServiceClient,
scan_setting: DataProfileScanSetting,
):
data_profile_spec = dataplex_v1.DataProfileSpec(
sampling_percent=scan_setting.sampling_percent,
row_filter=scan_setting.row_filter,
)
display_name = (
f"Data Profile Scan for {scan_setting.table_id} in {scan_setting.dataset_id}"
)
description = f"This is a Data Profile Scan for {scan_setting.project_id}.{scan_setting.dataset_id}.{scan_setting.table_id}. Created by dbt."
labels = {
"managed_by": "dbt",
}

if scan_setting.cron:
trigger = dataplex_v1.Trigger(
schedule=dataplex_v1.Trigger.Schedule(cron=scan_setting.cron)
)
else:
trigger = dataplex_v1.Trigger(on_demand=dataplex_v1.Trigger.OnDemand())
execution_spec = dataplex_v1.DataScan.ExecutionSpec(trigger=trigger)

if all(
scan.name != scan_setting.data_scan_name() for scan in client.list_data_scans(parent=scan_setting.parent())
):
data_scan = dataplex_v1.DataScan(
data=dataplex_v1.DataSource(
resource=f"//bigquery.googleapis.com/projects/{scan_setting.project_id}/datasets/{scan_setting.dataset_id}/tables/{scan_setting.table_id}"
),
data_profile_spec=data_profile_spec,
execution_spec=execution_spec,
display_name=display_name,
description=description,
labels=labels,
)
request = dataplex_v1.CreateDataScanRequest(
parent=scan_setting.parent(),
data_scan_id=scan_setting.scan_id,
data_scan=data_scan,
)
client.create_data_scan(request=request).result()
else:
request = dataplex_v1.GetDataScanRequest(
name=scan_setting.data_scan_name(),
)
data_scan = client.get_data_scan(request=request)

data_scan.data_profile_spec = data_profile_spec
data_scan.execution_spec = execution_spec
data_scan.display_name = display_name
data_scan.description = description
data_scan.labels = labels

update_mask = field_mask_pb2.FieldMask(
paths=[
"data_profile_spec",
"execution_spec",
"display_name",
"description",
"labels",
]
)
request = dataplex_v1.UpdateDataScanRequest(
data_scan=data_scan,
update_mask=update_mask,
)
client.update_data_scan(request=request).result()

@available
def create_or_update_data_profile_scan(self, config):
project_id = config.get("database")
dataset_id = config.get("schema")
table_id = config.get("name")

data_profile_config = config.get("config").get("data_profile_scan", {})

# Skip if data_profile_scan is not configured
if not data_profile_config:
return None

client = dataplex_v1.DataScanServiceClient()
scan_setting = DataProfileScanSetting(
location=data_profile_config["location"],
scan_id=data_profile_config.get(
"scan_id", self._generate_unique_scan_id(dataset_id, table_id)
),

project_id=project_id,
dataset_id=dataset_id,
table_id=table_id,
sampling_percent=data_profile_config.get("sampling_percent", None),
row_filter=data_profile_config.get("row_filter", None),
cron=data_profile_config.get("cron", None),
)

# Delete existing data profile scan if it is disabled
if not data_profile_config.get("enabled", True):
client.delete_data_scan(name=scan_setting.data_scan_name())
return None

self._create_or_update_data_profile_scan(client, scan_setting)

if not scan_setting.cron:
client.run_data_scan(
request=dataplex_v1.RunDataScanRequest(name=scan_setting.data_scan_name())
)

self._update_labels_with_data_profile_scan_labels(
project_id, dataset_id, table_id, scan_setting.location, scan_setting.scan_id
)

0 comments on commit 1a13aac

Please sign in to comment.