yujunjun/datahub
2
0
Fork 0
mirror of https://github.com/datahub-project/datahub.git synced 2025-12-16 12:38:13 +00:00
Code Issues Packages Projects Releases Wiki Activity

ingest(snowflake): remove email_as_user_identifier support (#13827)

Browse Source
This commit is contained in:
Aseem Bansal 2025-06-20 19:38:16 +05:30 committed by GitHub
parent b3a25d6fbd
commit f8c6db07d8
No known key found for this signature in database
GPG Key ID: B5690EEEBB952194
4 changed files with 7 additions and 18 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

9
metadata-ingestion/src/datahub/ingestion/source/snowflake/snowflake_config.py
View File

@ -154,14 +154,11 @@ class SnowflakeIdentifierConfig(
email_domain: Optional[str] = pydantic.Field( email_domain: Optional[str] = pydantic.Field(
default=None, default=None,
description="Email domain of your organization so users can be displayed on UI appropriately.", description="Email domain of your organization so users can be displayed on UI appropriately. This is used only if we cannot infer email ID.",
) )
email_as_user_identifier: bool = Field( _email_as_user_identifier = pydantic_removed_field(
default=True, "email_as_user_identifier",
description="Format user urns as an email, if the snowflake user's email is set. If `email_domain` is "
"provided, generates email addresses for snowflake users with unset emails, based on their "
"username.",
) )

9
metadata-ingestion/src/datahub/ingestion/source/snowflake/snowflake_utils.py
View File

@ -325,15 +325,10 @@ class SnowflakeIdentifierBuilder:
user_email: Optional[str], user_email: Optional[str],
) -> str: ) -> str:
if user_email: if user_email:
return self.snowflake_identifier( return self.snowflake_identifier(user_email)
user_email
if self.identifier_config.email_as_user_identifier is True
else user_email.split("@")[0]
)
return self.snowflake_identifier( return self.snowflake_identifier(
f"{user_name}@{self.identifier_config.email_domain}" f"{user_name}@{self.identifier_config.email_domain}"
if self.identifier_config.email_as_user_identifier is True if self.identifier_config.email_domain is not None
and self.identifier_config.email_domain is not None
else user_name else user_name
) )

1
metadata-ingestion/tests/integration/snowflake/test_snowflake.py
View File

@ -121,7 +121,6 @@ def test_snowflake_basic(pytestconfig, tmp_path, mock_time, mock_datahub_graph):
format_sql_queries=True, format_sql_queries=True,
validate_upstreams_against_patterns=False, validate_upstreams_against_patterns=False,
include_operational_stats=True, include_operational_stats=True,
email_as_user_identifier=True,
incremental_lineage=False, incremental_lineage=False,
start_time=datetime(2022, 6, 6, 0, 0, 0, 0).replace( start_time=datetime(2022, 6, 6, 0, 0, 0, 0).replace(
tzinfo=timezone.utc tzinfo=timezone.utc

6
metadata-ingestion/tests/integration/snowflake/test_snowflake_queries.py
View File

@ -36,7 +36,6 @@ def test_user_identifiers_email_as_identifier(snowflake_connect, tmp_path):
"username": "TST_USR", "username": "TST_USR",
"password": "TST_PWD", "password": "TST_PWD",
}, },
"email_as_user_identifier": True,
"email_domain": "example.com", "email_domain": "example.com",
}, },
PipelineContext("run-id"), PipelineContext("run-id"),
@ -63,7 +62,7 @@ def test_user_identifiers_email_as_identifier(snowflake_connect, tmp_path):
@patch("snowflake.connector.connect") @patch("snowflake.connector.connect")
def test_user_identifiers_username_as_identifier(snowflake_connect, tmp_path): def test_user_identifiers_user_email_as_identifier(snowflake_connect, tmp_path):
source = SnowflakeQueriesSource.create( source = SnowflakeQueriesSource.create(
{ {
"connection": { "connection": {
@ -71,13 +70,12 @@ def test_user_identifiers_username_as_identifier(snowflake_connect, tmp_path):
"username": "TST_USR", "username": "TST_USR",
"password": "TST_PWD", "password": "TST_PWD",
}, },
"email_as_user_identifier": False,
}, },
PipelineContext("run-id"), PipelineContext("run-id"),
) )
assert ( assert (
source.identifiers.get_user_identifier("username", "username@example.com") source.identifiers.get_user_identifier("username", "username@example.com")
== "username" == "username@example.com"
) )
assert source.identifiers.get_user_identifier("username", None) == "username" assert source.identifiers.get_user_identifier("username", None) == "username"