This is an automated email from the ASF dual-hosted git repository.
curth pushed a commit to branch main
in repository https://gitbox.apache.org/repos/asf/arrow-adbc.git
The following commit(s) were added to refs/heads/main by this push:
new 07243a4f2 feat(csharp/src/Drivers/BigQuery): use a default project ID
if one is not specified (#2471)
07243a4f2 is described below
commit 07243a4f228a3107a70cb4c760374714cc5b2062
Author: davidhcoe <[email protected]>
AuthorDate: Tue Jan 21 08:50:32 2025 -0500
feat(csharp/src/Drivers/BigQuery): use a default project ID if one is not
specified (#2471)
- No longer requires a project ID to be specified in the parameters
- If not specified, will use a default project ID (same as the Go
driver)
- Adds parameter to include the public project ID if desired
---------
Co-authored-by: David Coe <>
---
csharp/src/Drivers/BigQuery/BigQueryConnection.cs | 29 +++++++++++++++----
csharp/src/Drivers/BigQuery/BigQueryParameters.cs | 4 +++
csharp/src/Drivers/BigQuery/readme.md | 5 +++-
.../Drivers/BigQuery/BigQueryTestConfiguration.cs | 5 +++-
.../test/Drivers/BigQuery/BigQueryTestingUtils.cs | 10 +++++--
csharp/test/Drivers/BigQuery/DriverTests.cs | 33 ++++++++++++++++++++--
6 files changed, 73 insertions(+), 13 deletions(-)
diff --git a/csharp/src/Drivers/BigQuery/BigQueryConnection.cs
b/csharp/src/Drivers/BigQuery/BigQueryConnection.cs
index 191526399..a0c70c63f 100644
--- a/csharp/src/Drivers/BigQuery/BigQueryConnection.cs
+++ b/csharp/src/Drivers/BigQuery/BigQueryConnection.cs
@@ -41,11 +41,13 @@ namespace Apache.Arrow.Adbc.Drivers.BigQuery
readonly IReadOnlyDictionary<string, string> properties;
BigQueryClient? client;
GoogleCredential? credential;
+ bool includePublicProjectIds = false;
const string infoDriverName = "ADBC BigQuery Driver";
const string infoDriverVersion = "1.0.0";
const string infoVendorName = "BigQuery";
const string infoDriverArrowVersion = "1.0.0";
+ const string publicProjectId = "bigquery-public-data";
readonly AdbcInfoCode[] infoSupportedCodes = new[] {
AdbcInfoCode.DriverName,
@@ -81,8 +83,15 @@ namespace Apache.Arrow.Adbc.Drivers.BigQuery
// TODO: handle token expiration
+ // if the caller doesn't specify a projectId, use the default
if (!this.properties.TryGetValue(BigQueryParameters.ProjectId, out
projectId))
- throw new ArgumentException($"The
{BigQueryParameters.ProjectId} parameter is not present");
+ projectId = BigQueryConstants.DetectProjectId;
+
+ if
(this.properties.TryGetValue(BigQueryParameters.IncludePublicProjectId, out
string? result))
+ {
+ if (!string.IsNullOrEmpty(result))
+ includePublicProjectIds = Convert.ToBoolean(result);
+ }
if
(this.properties.TryGetValue(BigQueryParameters.AuthenticationType, out string?
newAuthenticationType))
{
@@ -308,11 +317,18 @@ namespace Apache.Arrow.Adbc.Drivers.BigQuery
if (catalogs != null)
{
- foreach (CloudProject catalog in catalogs)
+ List<string> projectIds = catalogs.Select(x =>
x.ProjectId).ToList();
+
+ if (this.includePublicProjectIds &&
!projectIds.Contains(publicProjectId))
+ projectIds.Add(publicProjectId);
+
+ projectIds.Sort();
+
+ foreach (string projectId in projectIds)
{
- if (Regex.IsMatch(catalog.ProjectId, catalogRegexp,
RegexOptions.IgnoreCase))
+ if (Regex.IsMatch(projectId, catalogRegexp,
RegexOptions.IgnoreCase))
{
- catalogNameBuilder.Append(catalog.ProjectId);
+ catalogNameBuilder.Append(projectId);
if (depth == GetObjectsDepth.Catalogs)
{
@@ -321,7 +337,7 @@ namespace Apache.Arrow.Adbc.Drivers.BigQuery
else
{
catalogDbSchemasValues.Add(GetDbSchemas(
- depth, catalog.ProjectId, dbSchemaPattern,
+ depth, projectId, dbSchemaPattern,
tableNamePattern, tableTypes,
columnNamePattern));
}
}
@@ -333,6 +349,7 @@ namespace Apache.Arrow.Adbc.Drivers.BigQuery
catalogNameBuilder.Build(),
catalogDbSchemasValues.BuildListArrayForType(new
StructType(StandardSchemas.DbSchemaSchema)),
};
+
StandardSchemas.GetObjectsSchema.Validate(dataArrays);
return dataArrays;
@@ -994,7 +1011,7 @@ namespace Apache.Arrow.Adbc.Drivers.BigQuery
{
if (this.credential == null)
{
- throw new InvalidOperationException();
+ throw new AdbcException("A credential must be set",
AdbcStatusCode.Unauthenticated);
}
if (this.client == null)
diff --git a/csharp/src/Drivers/BigQuery/BigQueryParameters.cs
b/csharp/src/Drivers/BigQuery/BigQueryParameters.cs
index 51272eb64..101a2dafa 100644
--- a/csharp/src/Drivers/BigQuery/BigQueryParameters.cs
+++ b/csharp/src/Drivers/BigQuery/BigQueryParameters.cs
@@ -36,6 +36,7 @@ namespace Apache.Arrow.Adbc.Drivers.BigQuery
public const string IncludeConstraintsWithGetObjects =
"adbc.bigquery.include_constraints_getobjects";
public const string GetQueryResultsOptionsTimeoutMinutes =
"adbc.bigquery.get_query_results_options.timeout";
public const string MaxFetchConcurrency =
"adbc.bigquery.max_fetch_concurrency";
+ public const string IncludePublicProjectId =
"adbc.bigquery.include_public_project_id";
}
/// <summary>
@@ -47,5 +48,8 @@ namespace Apache.Arrow.Adbc.Drivers.BigQuery
public const string ServiceAccountAuthenticationType = "service";
public const string TokenEndpoint =
"https://accounts.google.com/o/oauth2/token";;
public const string TreatLargeDecimalAsString = "true";
+
+ // default value per
https://pkg.go.dev/cloud.google.com/go/bigquery#section-readme
+ public const string DetectProjectId = "*detect-project-id*";
}
}
diff --git a/csharp/src/Drivers/BigQuery/readme.md
b/csharp/src/Drivers/BigQuery/readme.md
index 92af3ef1e..f7adb3958 100644
--- a/csharp/src/Drivers/BigQuery/readme.md
+++ b/csharp/src/Drivers/BigQuery/readme.md
@@ -64,7 +64,10 @@
https://cloud.google.com/dotnet/docs/reference/Google.Cloud.BigQuery.V2/latest/G
Optional. Sets the
[DestinationTable](https://cloud.google.com/dotnet/docs/reference/Google.Cloud.BigQuery.V2/latest/Google.Cloud.BigQuery.V2.QueryOptions#Google_Cloud_BigQuery_V2_QueryOptions_DestinationTable)
value of the QueryOptions if configured. Expects the format to be
`{projectId}.{datasetId}.{tableId}` to set the corresponding values in the
[TableReference](https://github.com/googleapis/google-api-dotnet-client/blob/6c415c73788b848711e47c6dd33c2f93c76faf9
[...]
**adbc.bigquery.project_id**<br>
- The [Project
ID](https://cloud.google.com/resource-manager/docs/creating-managing-projects)
used for accessing BigQuery.
+ The [Project
ID](https://cloud.google.com/resource-manager/docs/creating-managing-projects)
used for accessing BigQuery. If not specified, will default to detect the
projectIds the credentials have access to.
+
+**adbc.bigquery.include_public_project_id**<br>
+ Include the `bigquery-public-data` project ID with the
list of project IDs.
**adbc.bigquery.refresh_token**<br>
The refresh token used for when the generated OAuth
token expires. Required for `user` authentication.
diff --git a/csharp/test/Drivers/BigQuery/BigQueryTestConfiguration.cs
b/csharp/test/Drivers/BigQuery/BigQueryTestConfiguration.cs
index ffcae7cc0..17c228ac9 100644
--- a/csharp/test/Drivers/BigQuery/BigQueryTestConfiguration.cs
+++ b/csharp/test/Drivers/BigQuery/BigQueryTestConfiguration.cs
@@ -31,7 +31,7 @@ namespace Apache.Arrow.Adbc.Tests.Drivers.BigQuery
}
[JsonPropertyName("projectId")]
- public string ProjectId { get; set; } = string.Empty;
+ public string? ProjectId { get; set; }
[JsonPropertyName("clientId")]
public string ClientId { get; set; } = string.Empty;
@@ -57,6 +57,9 @@ namespace Apache.Arrow.Adbc.Tests.Drivers.BigQuery
[JsonPropertyName("includeTableConstraints")]
public bool IncludeTableConstraints { get; set; }
+ [JsonPropertyName("includePublicProjectId")]
+ public bool IncludePublicProjectId { get; set; } = false;
+
[JsonPropertyName("timeoutMinutes")]
public int? TimeoutMinutes { get; set; }
diff --git a/csharp/test/Drivers/BigQuery/BigQueryTestingUtils.cs
b/csharp/test/Drivers/BigQuery/BigQueryTestingUtils.cs
index 2bc6227ba..ff7e36139 100644
--- a/csharp/test/Drivers/BigQuery/BigQueryTestingUtils.cs
+++ b/csharp/test/Drivers/BigQuery/BigQueryTestingUtils.cs
@@ -51,10 +51,12 @@ namespace Apache.Arrow.Adbc.Tests.Drivers.BigQuery
/// <returns></returns>
internal static Dictionary<string, string>
GetBigQueryParameters(BigQueryTestConfiguration testConfiguration)
{
- Dictionary<string, string> parameters = new Dictionary<string,
string>
+ Dictionary<string, string> parameters = new Dictionary<string,
string>{};
+
+ if (!string.IsNullOrEmpty(testConfiguration.ProjectId))
{
- { BigQueryParameters.ProjectId, testConfiguration.ProjectId },
- };
+ parameters.Add(BigQueryParameters.ProjectId,
testConfiguration.ProjectId!);
+ }
if (!string.IsNullOrEmpty(testConfiguration.JsonCredential))
{
@@ -81,6 +83,8 @@ namespace Apache.Arrow.Adbc.Tests.Drivers.BigQuery
parameters.Add(BigQueryParameters.IncludeConstraintsWithGetObjects,
testConfiguration.IncludeTableConstraints.ToString());
+ parameters.Add(BigQueryParameters.IncludePublicProjectId,
testConfiguration.IncludePublicProjectId.ToString());
+
if
(!string.IsNullOrEmpty(testConfiguration.LargeResultsDestinationTable))
{
parameters.Add(BigQueryParameters.LargeResultsDestinationTable,
testConfiguration.LargeResultsDestinationTable);
diff --git a/csharp/test/Drivers/BigQuery/DriverTests.cs
b/csharp/test/Drivers/BigQuery/DriverTests.cs
index 579a2a83c..58db4db16 100644
--- a/csharp/test/Drivers/BigQuery/DriverTests.cs
+++ b/csharp/test/Drivers/BigQuery/DriverTests.cs
@@ -23,6 +23,7 @@ using Apache.Arrow.Adbc.Tests.Metadata;
using Apache.Arrow.Adbc.Tests.Xunit;
using Apache.Arrow.Ipc;
using Xunit;
+using Xunit.Abstractions;
namespace Apache.Arrow.Adbc.Tests.Drivers.BigQuery
{
@@ -37,12 +38,14 @@ namespace Apache.Arrow.Adbc.Tests.Drivers.BigQuery
public class DriverTests
{
BigQueryTestConfiguration _testConfiguration;
+ readonly ITestOutputHelper? outputHelper;
- public DriverTests()
+ public DriverTests(ITestOutputHelper? outputHelper)
{
+ this.outputHelper = outputHelper;
+
Skip.IfNot(Utils.CanExecuteTestConfig(BigQueryTestingUtils.BIGQUERY_TEST_CONFIG_VARIABLE));
_testConfiguration =
Utils.LoadTestConfiguration<BigQueryTestConfiguration>(BigQueryTestingUtils.BIGQUERY_TEST_CONFIG_VARIABLE);
-
}
/// <summary>
@@ -98,6 +101,32 @@ namespace Apache.Arrow.Adbc.Tests.Drivers.BigQuery
}
}
+ /// <summary>
+ /// Validates if the driver can call GetObjects.
+ /// </summary>
+ [SkippableFact, Order(3)]
+ public void CanGetObjectsAllCatalogs()
+ {
+ AdbcConnection adbcConnection =
BigQueryTestingUtils.GetBigQueryAdbcConnection(_testConfiguration);
+
+ IArrowArrayStream stream = adbcConnection.GetObjects(
+ depth: AdbcConnection.GetObjectsDepth.Catalogs,
+ catalogPattern: null,
+ dbSchemaPattern: null,
+ tableNamePattern: null,
+ tableTypes: BigQueryTableTypes.TableTypes,
+ columnNamePattern: null);
+
+ RecordBatch recordBatch = stream.ReadNextRecordBatchAsync().Result;
+
+ List<AdbcCatalog> catalogs =
GetObjectsParser.ParseCatalog(recordBatch, null);
+
+ foreach (AdbcCatalog ct in catalogs)
+ {
+ this.outputHelper?.WriteLine(ct.Name);
+ }
+ }
+
/// <summary>
/// Validates if the driver can call GetObjects.
/// </summary>
