HierarchyId GetHashCode Method returns a limited set of values

[gmw105]

Bug description

Having upgraded my project from .NET Framework to .Net 8, and moving to EFCore, I encountered a drastic downturn in performance of client side queries and dictionaries involving HierarchyId.
On investigation I found that the HierarchyId GetHashCode method is providing hash values of only 0-255 (i.e. it is only using the first byte of the 32-bit integer).
This method refers to the GetHashCode method from Microsoft.SqlServer.Types.SqlHierarchyId, so the problem originates there.

The code snippet demonstrates the values for a small test case (all output values are between 0-255).

Your code

using Microsoft.EntityFrameworkCore;

HierarchyId[] hidArray =
    [
        HierarchyId.GetRoot()
        , HierarchyId.Parse(@"/1/1/")
        , HierarchyId.Parse(@"/1/2/3/1/")
        , HierarchyId.Parse(@"/2/19/11/108/43/")
        , HierarchyId.Parse(@"/44/12.5/721.23/1.1/")
        , HierarchyId.Parse(@"/-32/19.1/93485/234/")
        , HierarchyId.Parse(@"/407/")
        , HierarchyId.Parse(@"/999999/1/1/1/")
        , HierarchyId.Parse(@"/12/11/10/9/8/7/6/5/4/3/2/1/")
        , HierarchyId.Parse(@"/-121/921873/564/2343/123/56/")
        , HierarchyId.Parse(@"/99/98/97/")
        , HierarchyId.Parse(@"/13/12/11/10/")
        , HierarchyId.Parse(@"/2/4/6/8/")
        , HierarchyId.Parse(@"/3/5/7/9/")
    ];


foreach(HierarchyId hid in hidArray)
{
    int hc = hid.GetHashCode();
    Console.WriteLine(hc.ToString());
}

Stack traces

Verbose output

EF Core version

9.0.9

Database provider

No response

Target framework

No response

Operating system

No response

IDE

No response

[roji]

This method refers to the GetHashCode method from Microsoft.SqlServer.Types.SqlHierarchyId, so the problem originates there.

If the problem is in Microsoft.SqlServer.Types.SqlHierarchyId, then unless I'm mistaken that's not part of EF, is it? Shouldn't this issue be moved to the sqlclient repo?

[gmw105]

Thanks for your reply. My apologies if I have posted in the wrong place - I am a bit lost on where the Microsoft.SqlServer.Types.SqlHierarchyId class lives. I can't see it in
SqlClient
, but perhaps I am missing it?

[roji]

Transferred this from the efcore repo to the SqlClient repo.

[cheenamalhotra]

/triage

[github-actions]

🔍 Triage Summary (on-demand re-triage)

Check	Result
Issue type	Bug
Environment	⚠️ Missing: .NET target framework, SQL Server version, Operating system, Database provider
Area	Area\Netcore
Duplicates	None found
Regression	Likely regression from .NET Framework — EF Core's HierarchyId on .NET delegates to Microsoft.SqlServer.Types.SqlHierarchyId.GetHashCode(), while the legacy SqlHierarchyId on .NET Framework used a different hash computation

Analysis

HierarchyId.GetHashCode() in the EF Core SqlServer HierarchyId wrapper delegates to the underlying Microsoft.SqlServer.Types.SqlHierarchyId.GetHashCode(), which appears to only consume the first byte of the internal binary representation, producing values in the range 0–255. This severely degrades performance of hash-based collections (dictionaries, hash sets) when using HierarchyId keys. While the root cause is in Microsoft.SqlServer.Types.SqlHierarchyId, the EF Core / SqlClient HierarchyId wrapper could override GetHashCode() with a proper implementation (e.g., using HashCode.Combine over the full binary value). This is a P2 issue: no data loss, but a significant performance regression for users migrating from .NET Framework to .NET 8+.

Next Steps

• @gmw105, please provide the following missing details so the team can fully characterize the environment:
  • .NET target framework (e.g., net8.0)
  • SQL Server version (e.g., SQL Server 2019)
  • Operating system (e.g., Windows 11, Ubuntu 22.04)
  • Database provider (e.g., Microsoft.EntityFrameworkCore.SqlServer and its version)
• Assign to Copilot coding agent to investigate the HierarchyId.GetHashCode() implementation in the Microsoft.SqlServer.Server project within this repo and the EF Core SqlServer HierarchyId type wrapper, to implement a proper hash combining the full binary representation of the HierarchyId value — for example using HashCode.Combine or SequenceHashCode over all bytes of the underlying value.
• Since the immediate wrapper lives in EF Core (Microsoft.EntityFrameworkCore.SqlServer), a parallel issue in dotnet/efcore may be warranted if the fix cannot be applied here.

Note: This triage summary is auto-generated by an AI agent. The analysis and suggestions above have not been verified by a human maintainer. Please treat as preliminary guidance only.

Generated by SqlClient Issue Auto-Triage for issue #3636 · 39.8 AIC · ⌖ 9.03 AIC · ⊞ 8.4K · ◷