GH-301: [C#] Use an index lookup for O(1) field index access

[vthemelis]

Closes #301.

Ports the optimization from the closed PR at apache/arrow#44633 into the new .NET-specific repository.

The original PR was closed on November 18, 2025 with the note that the C# implementation had moved to a new repository.

This version keeps the current arrow-dotnet behavior intact:

• GetFieldIndex(..., comparer: null) and the default path now use a cached CurrentCulture index lookup for the common case.
• Missing fields still return -1.
• Duplicate field names still return the first match.
• Non-default comparers still fall back to the existing linear scan.

I also added dedicated schema tests covering:

• null, Ordinal, OrdinalIgnoreCase, and CurrentCulture comparers
• duplicate-name lookup returning the first match
• missing-name behavior for each comparer

Local verification:

• dotnet build test/Apache.Arrow.Tests/Apache.Arrow.Tests.csproj
• DOTNET_ROLL_FORWARD=Major DOTNET_ROLL_FORWARD_TO_PRERELEASE=1 dotnet test test/Apache.Arrow.Tests/Apache.Arrow.Tests.csproj --no-restore --logger 'console;verbosity=minimal'

[adamreeve]

This looks good to me thanks @vthemelis

From apache/arrow#44633 there were a couple of issues raised:

• This doesn't support lookup of multiple indexes for the same name. I think this is OK, the PR just speeds up the existing API without changing behaviour (the first matching index is returned). If there's a need to support lookup with multiple results that could be a separate change.
• Use of StringComparer.CurrentCulture. I think it's correct to use this here to avoid changing behaviour. If we want to switch to the Ordinal comparer being the default, that should be a separate change.