Let `create-catalogs-schemas` reuse `MigrateGrants` so that it applies group renaming #2955

JCZuurmond · 2024-10-14T15:39:37Z

Changes

Let create-catalogs-schemas reuse MigrateGrants so that it applies group renaming and eliminate duplicate code.

Linked issues

Resolves #2934
Resolves #2932

Functionality

modified existing command: databricks labs ucx create-catalogs-schemas
modified existing workflow: migrate-tables (*functionally no change)

Tests

manually tested
added unit tests
added integration tests

JCZuurmond · 2024-10-14T15:43:29Z

src/databricks/labs/ucx/hive_metastore/catalog_schema.py

+        catalogs_existing, schemas_existing = self._catalogs_schemas_from_unity_catalog()
+        catalogs, schemas = self._catalogs_schemas_from_table_mapping()
+        for src_schema, dst_catalog in catalogs:
+            if dst_catalog in catalogs_existing:


@nfx and @FastLee : This introduces the question on "do we want to skip applying ACLs when the catalog exists?"

Before this PR:

ACLs is still applied to an existing catalog if a new schema will be created within that catalog

ACLs will be skipped for an existing schema.

@FastLee to give final opinion. i think we should apply to existing catalog, but i don't feel strongly about it

I am leaning to applying all ACLs present also if the catalog exist. Then, provide an option to the user to skip the ACLs alltogether. So either all or no ACLs are applied

I would skip.
Let's just alert.

Skip what exactly? The ones for the catalog if the catalog exists, the ones for the schema if the schema exits, the ones for the catalog if the schema exits? All ACLs if any schema or catalog exists?

Currently implemented that all legacy grants are migrated to UC and created an issue to make this part optional.

nfx · 2024-10-15T13:29:10Z

src/databricks/labs/ucx/hive_metastore/objects.py

+
+
+@dataclass(frozen=True)
+class Catalog:


~~this data structure seems to be used only in one place. why do we need to create whole new modules for those? also, it really is tuple[str,str] = ("CATALOG", name)~~

might be a good idea to co-locate next to Table in src/databricks/labs/ucx/hive_metastore/tables.py. make sure to implement only the required interface methods.

Move this to a separate module to avoid circular imports. After introducing a protocol in the grants module (see comment below) this was not an issue anymore. Decided to move this to the catalog_schema module as this is closer to the pattern that we more often use for hive_metastore/ modules, i.e. the object being crawled in a module is represented by a dataclass defined in the same module.

src/databricks/labs/ucx/hive_metastore/catalog_schema.py

nfx · 2024-10-15T13:32:29Z

src/databricks/labs/ucx/hive_metastore/grants.py

+        match self.action_type:
+            case "OWN":
+                return 1
+            case _:
+                return 0


Suggested change

match self.action_type:

case "OWN":

return 1

case _:

return 0

if self.action_type == "OWN":

return 1

return 0

why do we need all this syntactic sugar to replace a simple if?

Because I like to be fancy and show off 💁‍♂️

It is copied from:

ucx/src/databricks/labs/ucx/installer/workflows.py

Lines 766 to 776 in d3d08f9

@staticmethod

def _library_dep_order(library: str):

match library:

case library if 'sdk' in library:

return 0

case library if 'blueprint' in library:

return 1

case library if 'sqlglot' in library:

return 2

case _:

return 3

.

Though, with just two values 0 and 1 it maybe does not matter

src/databricks/labs/ucx/hive_metastore/grants.py

nfx · 2024-10-15T13:35:55Z

src/databricks/labs/ucx/hive_metastore/grants.py

@@ -765,16 +782,14 @@ def _grants(self) -> list[Grant]:
                grants.append(grant)
        return grants

-    def _match_grants(self, table: Table) -> list[Grant]:
+    def _match_grants(self, src: Catalog | Schema | Table) -> list[Grant]:


same protocol here. https://typing.readthedocs.io/en/latest/spec/protocol.html#protocols

nfx · 2024-10-15T13:43:15Z

tests/integration/hive_metastore/test_catalog_schema.py

+@retried(on=[NotFound], timeout=timedelta(seconds=20))
+def get_catalog(ws: WorkspaceClient, name: str) -> CatalogInfo:
+    return ws.catalogs.get(name)
+
+
+@retried(on=[NotFound], timeout=timedelta(seconds=20))
+def get_schema(ws: WorkspaceClient, full_name: str) -> SchemaInfo:
+    return ws.schemas.get(full_name)
+
+
+@retried(on=[NotFound], timeout=timedelta(seconds=20))
+def get_schema_permissions_list(ws: WorkspaceClient, full_name: str) -> PermissionsList:
+    return ws.grants.get(SecurableType.SCHEMA, full_name)


move this logic down to CatalogSchema class, so that you handle eventual consistency in production code, not in the test code. remember - CREATE-READ-READ is the pattern used in Terraform to handle these situations.

e.g. this makes the wrong thing more robust :)

See the changes to the CatalogSchema class.

Note that I went for a (READ-)CREATE-READ pattern:

READ before CREATE to check if resource already exists.

One READ after create because all settings can be set during the CREATE which is not always the case for Terraform resources.

Also note that I kept get_schema_permissions_list as this uses a different API for CREATE (with SQL) and READ (with SDK/RestAPI). For now, I do not think it is worth it to implement a mapping between these two APIs to introduce the CREATE-READ pattern.

https://github.com/databrickslabs/ucx/pull/2955/files/a9915fb6bb5cc24a52d20f3702643afd8824cdc8..e3704aa28f6232b24489cfd62d7b80e69f6f8515

JCZuurmond

See comments

JCZuurmond · 2024-10-15T13:36:47Z

src/databricks/labs/ucx/hive_metastore/objects.py

@@ -0,0 +1,36 @@
+from dataclasses import dataclass


@nfx: Have a look at this module.

Notes:

Needed to be a separate module to avoid circular imports

Maybe we want to introduce a parent object securable object (later), to have a common interface for securable objects

Note some inconsistencies across this code base and including the sdk:

SchemaInfo(catalog_name="catalog") vs Schema(catalog="catalog"). I choose the later to be consistent with the already existing Table class

Should we have a Schema or Database?

When do we use full_name and when key?

More?

Needed to be a separate module to avoid circular imports

Not a problem anymore after using a Protocol, #2955 (comment)

Added the above as code comments

JCZuurmond · 2024-10-15T13:40:29Z

src/databricks/labs/ucx/hive_metastore/mapping.py

+    @property
+    def as_uc_table(self) -> Table:
+        return Table(self.catalog_name, self.dst_schema, self.dst_table)
+
    @property
    def as_uc_table_key(self) -> str:


Not for this PR, but would try to avoid using this key and use the Table data class instead

JCZuurmond · 2024-10-15T13:41:14Z

src/databricks/labs/ucx/hive_metastore/tables.py

@@ -53,8 +53,8 @@ class Table:  # pylint: disable=too-many-public-methods
    catalog: str
    database: str
    name: str
-    object_type: str
-    table_format: str
+    object_type: str | None = None


Note change on this data type to support Rule.as_uc_table

this will cause table migration issues. can we get without this change?

JCZuurmond · 2024-10-15T13:42:14Z

tests/integration/hive_metastore/test_catalog_schema.py

-        schema_name=src_schema.name,
-        external_csv=make_mounted_location,
-    )
-    dst_catalog = make_catalog()


@FastLee : I removed this to test create a new catalog. Please review if the test is correct after my changes

JCZuurmond · 2024-10-15T13:43:34Z

tests/unit/hive_metastore/test_catalog_schema.py

+
+
+def prepare_test(ws, backend: MockBackend | None = None) -> CatalogSchema:  # pylint: disable=too-complex
+    """Prepare tests with the following setup:


Did a clean up of the prepare_tests to verify it is testing what I expect it does

JCZuurmond · 2024-10-15T13:45:13Z

src/databricks/labs/ucx/hive_metastore/grants.py

@@ -737,20 +752,22 @@ def __init__(
        self._group_manager = group_manager
        self._grant_loaders = grant_loaders

-    def apply(self, src: Table, uc_table_key: str) -> bool:
+    def apply(self, src: Catalog | Schema | Table, dst: Catalog | Schema | Table) -> bool:


This type change allows to match on a Catalog, Schema and Table not just table key.

github-actions · 2024-10-15T14:12:34Z

✅ 52/52 passed, 4 skipped, 1h55m4s total

_{Running from acceptance #6813}

src/databricks/labs/ucx/hive_metastore/tables.py

nfx

one last change and we're good to merge

src/databricks/labs/ucx/hive_metastore/catalog_schema.py

nfx · 2024-10-16T12:11:38Z

src/databricks/labs/ucx/hive_metastore/tables.py

@@ -53,8 +53,8 @@ class Table:  # pylint: disable=too-many-public-methods
    catalog: str
    database: str
    name: str
-    object_type: str
-    table_format: str
+    object_type: str | None = None


this will cause table migration issues. can we get without this change?

#2975

#2979

…ests

* Added DBFS Root resolution when HMS Federation is enabled ([#2947](#2947)). This commit introduces a DBFS resolver for use with HMS (Hive Metastore) federation, enabling accurate resolution of DBFS root locations when HMS federation is enabled. A new `_resolve_dbfs_root()` class method is added to the `MountsCrawler` class, and a boolean argument `enable_hms_federation` is included in the `MountsCrawler` constructor, providing better handling of federation functionality. The commit also adds a test function, `test_resolve_dbfs_root_in_hms_federation`, to validate the resolution of DBFS roots with HMS federation. The test covers special cases, such as the `/user/hive/metastore` path, and utilizes `LocationTrie` for more accurate location guessing. These changes aim to improve the overall DBFS root resolution when using HMS federation. * Added `jax-jumpy` to known list ([#2959](#2959)). In this release, we have added the `jax-jumpy` package to the list of known packages in our system. `jax-jumpy` is a Python-based numerical computation library, which includes modules such as `jumpy`, `jumpy._base_fns`, `jumpy.core`, `jumpy.lax`, `jumpy.numpy`, `jumpy.numpy._factory_fns`, `jumpy.numpy._transform_data`, `jumpy.numpy._types`, `jumpy.numpy.linalg`, `jumpy.ops`, and `jumpy.random`. These modules are now recognized by our system, which partially resolves issue [#1931](#1931), which may have been caused by the integration of the `jax-jumpy` package. Engineers can now utilize the capabilities of this library in their numerical computations. * Added `joblibspark` to known list ([#2960](#2960)). In this release, we have added support for the `joblibspark` library in our system by updating the `known.json` file, which keeps track of various libraries and their associated components. This change is a part of the resolution for issue [#1931](#1931) and includes new elements such as `doc`, `doc.conf`, `joblibspark`, `joblibspark.backend`, and `joblibspark.utils`. These additions enable the system to recognize and manage the new components related to `joblibspark`, allowing for improved compatibility and functionality. * Added `jsonpatch` to known list ([#2969](#2969)). In this release, we have added `jsonpatch` to the list of known libraries in the `known.json` file. Jsonpatch is a library used for applying JSON patches, which allow for partial updates to a JSON document. By including jsonpatch in the known list, developers can now easily utilize its functionality for JSON patching, and any necessary dependencies will be handled automatically. This change partially addresses issue [#1931](#1931), which may have been caused by the use or integration of jsonpatch. We encourage developers to take advantage of this new addition to enhance their code and efficiently make partial updates to JSON documents. * Added `langchain-community` to known list ([#2970](#2970)). A new entry for `langchain-community` has been added to the configuration file for known language chain components in this release. This entry includes several sub-components such as 'langchain_community.agents', 'langchain_community.callbacks', 'langchain_community.chat_loaders', 'langchain_community.chat_message_histories', 'langchain_community.chat_models', 'langchain_community.cross_encoders', 'langchain_community.docstore', 'langchain_community.document_compressors', 'langchain_community.document_loaders', 'langchain_community.document_transformers', 'langchain_community.embeddings', 'langchain_community.example_selectors', 'langchain_community.graph_vectorstores', 'langchain_community.graphs', 'langchain_community.indexes', 'langchain_community.llms', 'langchain_community.memory', 'langchain_community.output_parsers', 'langchain_community.query_constructors', 'langchain_community.retrievers', 'langchain_community.storage', 'langchain_community.tools', 'langchain_community.utilities', and 'langchain_community.utils'. Currently, these sub-components are empty and have no additional configuration or code. This change partially resolves issue [#1931](#1931), but the specifics of the issue and how these components will be used are still unclear. * Added `langcodes` to known list ([#2971](#2971)). A new `langcodes` library has been added to the project, addressing part of issue [#1931](#1931). This library includes several modules that provide functionalities related to language codes and their manipulation, including `langcodes`, `langcodes.build_data`, `langcodes.data_dicts`, `langcodes.language_distance`, `langcodes.language_lists`, `langcodes.registry_parser`, `langcodes.tag_parser`, and `langcodes.util`. Additionally, the memory-efficient trie (prefix tree) data structure library, `marisa-trie`, has been included in the known list. It is important to note that no existing functionality has been altered in this commit. * Addressing Ownership Conflict when creating catalog/schemas ([#2956](#2956)). This release introduces new functionality to handle ownership conflicts during catalog/schema creation in our open-source library. The `_apply_from_legacy_table_acls` method has been enhanced with two loops to address non-own grants and own grants separately. This ensures proper handling of ownership conflicts by generating and executing UC grant SQL for each grant type, with appropriate exceptions. Additionally, a new helper function, `this_type_and_key()`, has been added to improve code readability. The release also introduces new methods, GrantsCrawler and Rule, in the hive_metastore package of the labs.ucx module, responsible for populating views and mapping source and destination objects. The test_catalog_schema.py file has been updated to include tests for creating catalogs and schemas with legacy ACLs, utilizing the new Rule method and GrantsCrawler. Issue [#2932](#2932) has been addressed with these changes, which include adding new methods and updating existing tests for hive_metastore. * Clarify `skip` and `unskip` commands work on views ([#2962](#2962)). In this release, the `skip` and `unskip` commands in the databricks labs UCX tool have been updated to clarify their functionality on views and to make it more explicit with the addition of the `--view` flag. These commands allow users to skip or unskip certain schemas, tables, or views during the table migration process. This is useful for temporarily disabling migration of a particular schema, table, or view. Unit tests have been added to ensure the correct behavior of these commands when working with views. Two new methods have been added to test the behavior of the `unskip` command when a schema or table is specified, and two additional methods test the behavior of the `unskip` command when a view or no schema is specified. Finally, two methods test that an error message is logged when both the `--table` and `--view` flags are specified. * Fixed issue with migrating MANAGED hive_metastore table to UC ([#2928](#2928)). This commit addresses an issue with migrating Hive Metastore (HMS) MANAGED tables to Unity Catalog (UC) as EXTERNAL, where deleting a MANAGED table can result in data loss. To prevent this, a new option `CONVERT_TO_EXTERNAL` has been added to the `migrate_tables` method for migrating managed tables to UC as external, ensuring that the HMS managed table is converted to an external table in HMS and UC, and protecting against data loss when deleting a managed table that has been migrated to UC as external. Additionally, new caching properties have been added for better performance, and existing methods have been modified to handle the migration of managed tables to UC as external. Tests, including unit and integration tests, have been added to ensure the proper functioning of these changes. It is important to note that changing MANAGED tables to EXTERNAL can have potential consequences on regulatory data cleanup, and the impact of this change should be carefully validated for existing workloads. * Let `create-catalogs-schemas` reuse `MigrateGrants` so that it applies group renaming ([#2955](#2955)). The `create-catalogs-schemas` command in the `databricks labs ucx` package has been enhanced to reuse the `MigrateGrants` function, enabling group renaming and eliminating redundant code. The `migrate-tables` workflow remains functionally the same. Changes include modifying the `CatalogSchema` class to accept a `migrate_grants` argument, introducing new `Catalog` and `Schema` dataclasses, and updating various methods in the `hive_metastore` module. Unit and integration tests have been added and manually verified to ensure proper functionality. The `MigrateGrants` class has been updated to accept two `SecurableObject` arguments and sort matched grants. The `from_src_dst` function in `mapping.py` now includes a new `as_uc_table` method and updates to `as_uc_table_key`. Addressing issues [#2934](#2934), [#2932](#2932), and [#2955](#2955), the changes also include a new `key` property for the `tables.py` file, and updates to the `test_create_catalogs_schemas` and `test_migrate_tables` test functions. * Updated sqlglot requirement from <25.25,>=25.5.0 to >=25.5.0,<25.26 ([#2968](#2968)). A update has been made to the sqlglot requirement in the pyproject.toml file, changing the version range from allowing versions 25.5.0 and later, but less than 25.25, to a range that allows for versions 25.5.0 and later, but less than 25.26. This change was implemented to allow for the latest version of sqlglot compatible with the project's requirements. The commit message includes a detailed changelog for sqlglot version 25.25.0, highlighting breaking changes, new features, bug fixes, and refactors. The commits included in the pull request are also outlined in the message, providing a comprehensive overview of the updates made. * Use `LocationTrie` to infer a list of UC external locations ([#2965](#2965)). This pull request introduces the `LocationTrie` class to improve the efficiency of determining external locations in UC storage. The `LocationTrie` class, implemented as a dataclass, maintains a hierarchical tree structure to store location paths and offers methods for inserting, finding, and checking if a table exists in the trie. It also refactors the existing `_parse_location` method and adds the `_parse_url` method to handle specific cases of JDBC URLs. The `find` method and `_external_locations` method are updated to use the new `LocationTrie` class, ensuring a more efficient way of determining overlapping storage prefixes. Additionally, the test file for external locations has been modified to include new URL formats and correct previously incorrect ones, enhancing compatibility with various URL formats. Overall, these changes are instrumental in preparing the codebase for future federated SQL connections by optimizing the determination of external locations and improving compatibility with diverse URL formats. * Warn when table has column without no name during table migration ([#2984](#2984)). In this release, we have introduced changes to the table migration feature in the databricks project to address issue [#2891](#2891). We have renamed the `_migrate_table` method in the `TableMigration` class to `_safe_migrate_table` and added a new `_migrate_table` method that includes a try-except block to catch a specific Spark AnalysisException caused by an invalid column name. If this exception is caught, a warning is logged, and the function returns False. This change ensures that the migration process continues even when it encounters a table with a column without a name, generating a warning instead of raising an error. Additionally, we have added unit tests to verify the new functionality, including the introduction of a new mock backend `MockBackendWithGeneralException` to raise a general exception for testing purposes and a new test case `test_migrate_tables_handles_table_with_empty_column` to validate the warning message generated when encountering a table with an empty column during migration. Dependency updates: * Updated sqlglot requirement from <25.25,>=25.5.0 to >=25.5.0,<25.26 ([#2968](#2968)).

JCZuurmond added bug Something isn't working migrate/access-control Access Control to things labels Oct 14, 2024

JCZuurmond self-assigned this Oct 14, 2024

JCZuurmond commented Oct 14, 2024

View reviewed changes

nfx requested a review from FastLee October 14, 2024 15:54

JCZuurmond force-pushed the fix/reuse-migrate-grants-in-create-catalogs-schemas branch 2 times, most recently from c634d3f to bb382f7 Compare October 15, 2024 13:23

nfx requested changes Oct 15, 2024

View reviewed changes

JCZuurmond commented Oct 15, 2024

View reviewed changes

JCZuurmond marked this pull request as ready for review October 15, 2024 13:45

JCZuurmond requested a review from a team as a code owner October 15, 2024 13:45

JCZuurmond temporarily deployed to account-admin October 15, 2024 13:45 — with GitHub Actions Inactive

This was referenced Oct 16, 2024

[TECH DEBT]: Change Schema kind to "SCHEMA" instead of "DATABASE" #2974

Open

[TECH DEBT]: Change securable object kind attribute to securable_type #2975

Open

JCZuurmond temporarily deployed to account-admin October 16, 2024 08:38 — with GitHub Actions Inactive

JCZuurmond commented Oct 16, 2024

View reviewed changes

src/databricks/labs/ucx/hive_metastore/tables.py Outdated Show resolved Hide resolved

JCZuurmond requested a review from nfx October 16, 2024 09:02

JCZuurmond had a problem deploying to account-admin October 16, 2024 09:06 — with GitHub Actions Error

JCZuurmond mentioned this pull request Oct 16, 2024

[TECH DEBT] Align Table.key with Databricks documentation #2979

Open

JCZuurmond temporarily deployed to account-admin October 16, 2024 09:30 — with GitHub Actions Inactive

nfx requested changes Oct 16, 2024

View reviewed changes

src/databricks/labs/ucx/hive_metastore/catalog_schema.py Outdated Show resolved Hide resolved

JCZuurmond temporarily deployed to account-admin October 16, 2024 10:00 — with GitHub Actions Inactive

JCZuurmond requested a review from nfx October 16, 2024 11:15

JCZuurmond force-pushed the fix/reuse-migrate-grants-in-create-catalogs-schemas branch from 7fda78f to c5069a9 Compare October 16, 2024 11:28

JCZuurmond temporarily deployed to account-admin October 16, 2024 11:28 — with GitHub Actions Inactive

nfx requested changes Oct 16, 2024

View reviewed changes

JCZuurmond had a problem deploying to account-admin October 16, 2024 13:51 — with GitHub Actions Error

JCZuurmond force-pushed the fix/reuse-migrate-grants-in-create-catalogs-schemas branch from 2cb3ca2 to 4a1aa19 Compare October 16, 2024 13:52

JCZuurmond temporarily deployed to account-admin October 16, 2024 13:52 — with GitHub Actions Inactive

JCZuurmond added 23 commits October 16, 2024 17:42

Format

0b82e41

Format

6920817

Add docstring to objects module

9bbe641

Refactor Schema dataclass to be consistent with code base

6092cf0

Add comment on order

867b09c

Move eventually consistent catalog logic into CatalogSchema

3edf9c2

Move eventually consistent schema logic into CatalogSchema

f61e30c

Catch DatabricksError when getting catalog or schema

09b339e

Introduce SecurableObject protocol

156357d

Add TODO

88391cf

Move Catalog and Schema to catalog_schema module

3e9a749

Add note on Schema kind being "DATABASE"

0497e9f

Add docstring for Catalog dataclass

32c23c9

Add note for kind

876c1ed

Add note for full name and key

abe2b46

Add TODO about kind

316ccbd

#2975

Add docstring for Schema

9b5a187

Add docstring for Catalog.name

ee76214

Move get_schema_permissions_list

83cdd0f

Change timeout to 30 seconds

63c1fc6

Update TODO with issue link

0e989fa

#2979

Move timeout to CatalogSchema constructor and set to None in unit t…

07208c6

…ests

Revert changes to Table dataclass

a7559ce

nfx force-pushed the fix/reuse-migrate-grants-in-create-catalogs-schemas branch from d9b2954 to a7559ce Compare October 16, 2024 15:44

nfx temporarily deployed to account-admin October 16, 2024 15:44 — with GitHub Actions Inactive

nfx merged commit fe06262 into main Oct 16, 2024
6 of 7 checks passed

nfx deleted the fix/reuse-migrate-grants-in-create-catalogs-schemas branch October 16, 2024 15:47

nfx mentioned this pull request Oct 16, 2024

Release v0.45.0 #2987

Merged

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Let `create-catalogs-schemas` reuse `MigrateGrants` so that it applies group renaming #2955

Let `create-catalogs-schemas` reuse `MigrateGrants` so that it applies group renaming #2955

JCZuurmond commented Oct 14, 2024 •

edited

Loading

JCZuurmond Oct 14, 2024 •

edited

Loading

nfx Oct 14, 2024

JCZuurmond Oct 14, 2024

FastLee Oct 14, 2024

JCZuurmond Oct 15, 2024

JCZuurmond Oct 15, 2024

nfx Oct 15, 2024

JCZuurmond Oct 16, 2024

nfx Oct 15, 2024

JCZuurmond Oct 16, 2024

nfx Oct 15, 2024

nfx Oct 15, 2024

JCZuurmond Oct 16, 2024

JCZuurmond Oct 16, 2024

JCZuurmond left a comment

JCZuurmond Oct 15, 2024

JCZuurmond Oct 16, 2024

JCZuurmond Oct 16, 2024

JCZuurmond Oct 15, 2024

JCZuurmond Oct 15, 2024

nfx Oct 16, 2024

JCZuurmond Oct 15, 2024

JCZuurmond Oct 15, 2024

JCZuurmond Oct 15, 2024

github-actions bot commented Oct 15, 2024 •

edited

Loading

nfx left a comment

nfx Oct 16, 2024

	@staticmethod
	def _library_dep_order(library: str):
	match library:
	case library if 'sdk' in library:
	return 0
	case library if 'blueprint' in library:
	return 1
	case library if 'sqlglot' in library:
	return 2
	case _:
	return 3



		def prepare_test(ws, backend: MockBackend \| None = None) -> CatalogSchema: # pylint: disable=too-complex
		"""Prepare tests with the following setup:

Let create-catalogs-schemas reuse MigrateGrants so that it applies group renaming #2955

Let create-catalogs-schemas reuse MigrateGrants so that it applies group renaming #2955

Conversation

JCZuurmond commented Oct 14, 2024 • edited Loading

Changes

Linked issues

Functionality

Tests

JCZuurmond Oct 14, 2024 • edited Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

JCZuurmond left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

github-actions bot commented Oct 15, 2024 • edited Loading

nfx left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Let `create-catalogs-schemas` reuse `MigrateGrants` so that it applies group renaming #2955

Let `create-catalogs-schemas` reuse `MigrateGrants` so that it applies group renaming #2955

JCZuurmond commented Oct 14, 2024 •

edited

Loading

JCZuurmond Oct 14, 2024 •

edited

Loading

github-actions bot commented Oct 15, 2024 •

edited

Loading