Fix MongoDB performance issue for case-insensitive SCIM filtering

[Copilot]

Problem

SCIM filtering on MongoDB was generating inefficient queries for case-insensitive string operations, causing severe performance degradation:

{
  "Attribute.SchemaAttributeId": "26d51050-4962-4348-a6cb-310c198eeee3",
  "$expr": {
    "$eq": [
      { "$toLower": { "$ifNull": ["$Attribute.ValueString", ""] } },
      "150017355"
    ]
  }
}

This query pattern forced MongoDB into full collection scans because $expr with $toLower and $ifNull functions prevent index usage, resulting in ~2.11 minutes execution time for 859k documents.

Solution

Implemented MongoDB-specific optimizations that replace problematic $expr queries with index-friendly regex patterns for case-insensitive string operations.

Key Changes

1. Optimized Expression Engine

• New MongoDbOptimizedExpressionExtensions class generates regex-based queries instead of $expr queries
• Uses anchored regex patterns (^value$) that MongoDB can optimize with indexes
• Handles all SCIM string operations: equality (eq), starts-with (sw), ends-with (ew), and contains (co)

2. Enhanced Index Strategy

• Added compound index: SchemaAttributeId_1_ValueString_1 for efficient attribute filtering
• Added case-insensitive collation index for ValueString field
• Maintains existing indexes for backward compatibility

3. Selective Optimization

• Only optimizes case-insensitive string operations (CaseExact = false)
• Preserves existing behavior for case-sensitive queries, non-string attributes, and complex expressions
• Transparent to API consumers - no application code changes required

Performance Impact

The optimization specifically targets the problematic query pattern:

• Before: $expr functions → Full collection scan → Poor performance
• After: Regex patterns → Index-supported queries → Significant improvement

Example Query Transformation

// Before (inefficient)
Expression.Call(Expression.Coalesce(property, Expression.Constant("")), "ToLower")

// After (optimized)  
var regex = new Regex($"^{Regex.Escape(value)}$", RegexOptions.IgnoreCase);
Expression.Call(regex, "IsMatch", property)

Testing

• Added comprehensive test suite with 5 new tests validating the optimization logic
• All existing SCIM tests (7) continue to pass, ensuring backward compatibility
• Tested various string operation scenarios: equality, starts-with, ends-with, contains

Backward Compatibility

✅ Fully backward compatible

• Case-sensitive queries unchanged
• Non-string attribute queries unchanged
• Complex attribute expressions unchanged
• All existing functionality preserved

The fix addresses the specific performance bottleneck reported in the issue while maintaining complete compatibility with existing SCIM implementations.

Original prompt

---

This section details on the original issue you should resolve

<issue_title>[SCIM] Performance issue filtering on MongoDb</issue_title>
<issue_description>Hello,

we are using SCIM on MongoDb database.

We facing issue calling this:

/Groups?filter=eventid eq "150017355"&startIndex=1&count=100&excludedAttributes=meta,schema,members

that is translated in:

{
"Attribute.SchemaAttributeId": "26d51050-4962-4348-a6cb-310c198eeee3",
"$expr": {
"$eq": [
{ "$toLower": { "$ifNull": ["$Attribute.ValueString", ""] } },
"150017355"
]
}
}

Because of $toLower and $ifNull, MongoDB cannot use the existing indexes and is forced into a full collection scan

Documents examined: ~859k
Documents returned: 9
Execution time: ~2.11 minutes

Can you please help us?

Regards.
Gabriele</issue_description>

Comments on the Issue (you are @copilot in this section)

Fixes #926

---

✨ Let Copilot coding agent set things up for you — coding agent works faster and does higher quality work when set up for your repo.