Return correct type when invoking GetFieldType and GetProviderSpecificFieldType for vector float32 column

[apoorvdeshmukh]

Description

Fixes #4104.
GetFieldType returned type from MetaType.ClassType
The fix follows similar pattern as GetValue to determine correct type for Vector metatype.
The fix has been extended to GetProviderSpecificFieldType which had similar issue.

Issues

#4104

Testing

Extended NativeVectorFloat32Tests with test coverage for above APIs.

Guidelines

Please review the contribution guidelines before submitting a pull request:

• Contributing
• Code of Conduct
• Best Practices
• Coding Style
• Review Process

[Copilot]

Pull request overview

Fixes SqlDataReader.GetFieldType() and GetProviderSpecificFieldType() to report the correct CLR type for vector(float32) columns (matching how GetValue() already materializes the value as SqlVector<float>), addressing issue #4104.

Changes:

• Update SqlDataReader to return typeof(SqlVector<float>) for SqlDbTypeExtensions.Vector (float32) in GetFieldTypeInternal.
• Apply the same correction to GetProviderSpecificFieldTypeInternal.
• Add a manual test covering both APIs for a vector(float32) column.

Reviewed changes

Copilot reviewed 2 out of 2 changed files in this pull request and generated 3 comments.

File	Description
src/Microsoft.Data.SqlClient/src/Microsoft/Data/SqlClient/SqlDataReader.cs	Special-cases Vector metadata so GetFieldType/GetProviderSpecificFieldType report SqlVector<float> instead of the metatype’s default byte[].
src/Microsoft.Data.SqlClient/tests/ManualTests/SQL/VectorTest/NativeVectorFloat32Tests.cs	Adds test coverage to validate the updated type reporting and alignment with GetValue().

[rhuijben]

Shouldn't this be fixed on the SqlBuffer layer. It already identifies SqlVector and does the deserialization.

Something like

diff --git a/src/Microsoft.Data.SqlClient/src/Microsoft/Data/SqlClient/SqlBuffer.cs b/src/Microsoft.Data.SqlClient/src/Microsoft/Data/SqlClient/SqlBuffer.cs
index dc1abf1c..04683c3c 100644
--- a/src/Microsoft.Data.SqlClient/src/Microsoft/Data/SqlClient/SqlBuffer.cs
+++ b/src/Microsoft.Data.SqlClient/src/Microsoft/Data/SqlClient/SqlBuffer.cs
@@ -1136,8 +1136,16 @@ namespace Microsoft.Data.SqlClient
                     case StorageType.String:
                         return String;
                     case StorageType.SqlBinary:
-                    case StorageType.Vector:
                         return ByteArray;
+                    case StorageType.Vector:
+                        var elementType = (MetaType.SqlVectorElementType)_value._vectorInfo._elementType;
+                        switch (elementType)
+                        {
+                            case MetaType.SqlVectorElementType.Float32:
+                                return GetSqlVector<float>();
+                            default:
+                                throw SQL.VectorTypeNotSupported(elementType.ToString());
+                        }
                     case StorageType.SqlCachedBuffer:
                         {
                             // If we have a CachedBuffer, it's because it's an XMLTYPE column
@@ -1213,6 +1221,15 @@ namespace Microsoft.Data.SqlClient
                     case StorageType.Json:
                         return typeof(SqlJson);
                         // Time Date DateTime2 and DateTimeOffset have no direct Sql type to contain them
+                    case StorageType.Vector:
+                        var elementType = (MetaType.SqlVectorElementType)_value._vectorInfo._elementType;
+                        switch (elementType)
+                        {
+                            case MetaType.SqlVectorElementType.Float32:
+                                return typeof(SqlVector<float>);
+                            default:
+                                throw SQL.VectorTypeNotSupported(elementType.ToString());
+                        }
                 }
             }
             else
@@ -1262,7 +1279,14 @@ namespace Microsoft.Data.SqlClient
                     case StorageType.Json:
                         return typeof(string);
                     case StorageType.Vector:
-                        return typeof(byte[]);
+                        var elementType = (MetaType.SqlVectorElementType)_value._vectorInfo._elementType;
+                        switch (elementType)
+                        {
+                            case MetaType.SqlVectorElementType.Float32:
+                                return typeof(SqlVector<float>);
+                            default:
+                                throw SQL.VectorTypeNotSupported(elementType.ToString());
+                        }
 #if NET
                     case StorageType.Time:
                         return typeof(TimeOnly);

I think this last hunk is the current issue. The type returned should be retrievable.

Not sure if some of this should be extracted to a helper function or something though. And more than just Float32 could be supported these days, like Half Precision Float Vectors

[apoorvdeshmukh]

Shouldn't this be fixed on the SqlBuffer layer. It already identifies SqlVector and does the deserialization.

Something like

diff --git a/src/Microsoft.Data.SqlClient/src/Microsoft/Data/SqlClient/SqlBuffer.cs b/src/Microsoft.Data.SqlClient/src/Microsoft/Data/SqlClient/SqlBuffer.cs
index dc1abf1c..04683c3c 100644
--- a/src/Microsoft.Data.SqlClient/src/Microsoft/Data/SqlClient/SqlBuffer.cs
+++ b/src/Microsoft.Data.SqlClient/src/Microsoft/Data/SqlClient/SqlBuffer.cs
@@ -1136,8 +1136,16 @@ namespace Microsoft.Data.SqlClient
                     case StorageType.String:
                         return String;
                     case StorageType.SqlBinary:
-                    case StorageType.Vector:
                         return ByteArray;
+                    case StorageType.Vector:
+                        var elementType = (MetaType.SqlVectorElementType)_value._vectorInfo._elementType;
+                        switch (elementType)
+                        {
+                            case MetaType.SqlVectorElementType.Float32:
+                                return GetSqlVector<float>();
+                            default:
+                                throw SQL.VectorTypeNotSupported(elementType.ToString());
+                        }
                     case StorageType.SqlCachedBuffer:
                         {
                             // If we have a CachedBuffer, it's because it's an XMLTYPE column
@@ -1213,6 +1221,15 @@ namespace Microsoft.Data.SqlClient
                     case StorageType.Json:
                         return typeof(SqlJson);
                         // Time Date DateTime2 and DateTimeOffset have no direct Sql type to contain them
+                    case StorageType.Vector:
+                        var elementType = (MetaType.SqlVectorElementType)_value._vectorInfo._elementType;
+                        switch (elementType)
+                        {
+                            case MetaType.SqlVectorElementType.Float32:
+                                return typeof(SqlVector<float>);
+                            default:
+                                throw SQL.VectorTypeNotSupported(elementType.ToString());
+                        }
                 }
             }
             else
@@ -1262,7 +1279,14 @@ namespace Microsoft.Data.SqlClient
                     case StorageType.Json:
                         return typeof(string);
                     case StorageType.Vector:
-                        return typeof(byte[]);
+                        var elementType = (MetaType.SqlVectorElementType)_value._vectorInfo._elementType;
+                        switch (elementType)
+                        {
+                            case MetaType.SqlVectorElementType.Float32:
+                                return typeof(SqlVector<float>);
+                            default:
+                                throw SQL.VectorTypeNotSupported(elementType.ToString());
+                        }
 #if NET
                     case StorageType.Time:
                         return typeof(TimeOnly);

I think this last hunk is the current issue. The type returned should be retrievable.

Not sure if some of this should be extracted to a helper function or something though. And more than just Float32 could be supported these days, like Half Precision Float Vectors

Shouldn't this be fixed on the SqlBuffer layer. It already identifies SqlVector and does the deserialization.

Something like

diff --git a/src/Microsoft.Data.SqlClient/src/Microsoft/Data/SqlClient/SqlBuffer.cs b/src/Microsoft.Data.SqlClient/src/Microsoft/Data/SqlClient/SqlBuffer.cs
index dc1abf1c..04683c3c 100644
--- a/src/Microsoft.Data.SqlClient/src/Microsoft/Data/SqlClient/SqlBuffer.cs
+++ b/src/Microsoft.Data.SqlClient/src/Microsoft/Data/SqlClient/SqlBuffer.cs
@@ -1136,8 +1136,16 @@ namespace Microsoft.Data.SqlClient
                     case StorageType.String:
                         return String;
                     case StorageType.SqlBinary:
-                    case StorageType.Vector:
                         return ByteArray;
+                    case StorageType.Vector:
+                        var elementType = (MetaType.SqlVectorElementType)_value._vectorInfo._elementType;
+                        switch (elementType)
+                        {
+                            case MetaType.SqlVectorElementType.Float32:
+                                return GetSqlVector<float>();
+                            default:
+                                throw SQL.VectorTypeNotSupported(elementType.ToString());
+                        }
                     case StorageType.SqlCachedBuffer:
                         {
                             // If we have a CachedBuffer, it's because it's an XMLTYPE column
@@ -1213,6 +1221,15 @@ namespace Microsoft.Data.SqlClient
                     case StorageType.Json:
                         return typeof(SqlJson);
                         // Time Date DateTime2 and DateTimeOffset have no direct Sql type to contain them
+                    case StorageType.Vector:
+                        var elementType = (MetaType.SqlVectorElementType)_value._vectorInfo._elementType;
+                        switch (elementType)
+                        {
+                            case MetaType.SqlVectorElementType.Float32:
+                                return typeof(SqlVector<float>);
+                            default:
+                                throw SQL.VectorTypeNotSupported(elementType.ToString());
+                        }
                 }
             }
             else
@@ -1262,7 +1279,14 @@ namespace Microsoft.Data.SqlClient
                     case StorageType.Json:
                         return typeof(string);
                     case StorageType.Vector:
-                        return typeof(byte[]);
+                        var elementType = (MetaType.SqlVectorElementType)_value._vectorInfo._elementType;
+                        switch (elementType)
+                        {
+                            case MetaType.SqlVectorElementType.Float32:
+                                return typeof(SqlVector<float>);
+                            default:
+                                throw SQL.VectorTypeNotSupported(elementType.ToString());
+                        }
 #if NET
                     case StorageType.Time:
                         return typeof(TimeOnly);

I think this last hunk is the current issue. The type returned should be retrievable.

Not sure if some of this should be extracted to a helper function or something though. And more than just Float32 could be supported these days, like Half Precision Float Vectors

Thanks for the detailed analysis @rhuijben! You're right that SqlBuffer.GetTypeFromStorageType(isSqlType: false) returning typeof(byte[]) for StorageType.Vector is inconsistent with how other APIs surface vector data. However, GetFieldType()/GetProviderSpecificFieldType() work from column metadata (available before Read()), not from SqlBuffer (populated after Read()), so the SqlDataReader fix is needed regardless.

And yes, support for Half precision vectors is coming soon in SqlClient. :)

[Copilot]

Pull request overview

Copilot reviewed 2 out of 2 changed files in this pull request and generated 1 comment.

[codecov]

Codecov Report

❌ Patch coverage is 93.75000% with 1 line in your changes missing coverage. Please review.
✅ Project coverage is 66.55%. Comparing base (60d4b92) to head (547705d).
⚠️ Report is 19 commits behind head on main.

Files with missing lines	Patch %	Lines
...ient/src/Microsoft/Data/SqlClient/SqlDataReader.cs	93.75%	1 Missing ⚠️

❗ There is a different number of reports uploaded between BASE (60d4b92) and HEAD (547705d). Click for more details.

HEAD has 1 upload less than BASE

Flag	BASE (60d4b92)	HEAD (547705d)
CI-SqlClient	1	0

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #4105      +/-   ##
==========================================
- Coverage   73.22%   66.55%   -6.67%     
==========================================
  Files         280      274       -6     
  Lines       43000    65798   +22798     
==========================================
+ Hits        31486    43794   +12308     
- Misses      11514    22004   +10490     

Flag	Coverage Δ
CI-SqlClient	?
PR-SqlClient-Project	66.55% <93.75%> (?)

Flags with carried forward coverage won't be shown. Click here to find out more.

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.