From 7a90b46c3b8eb43ff17fd699cc6d56f7d7b84fb3 Mon Sep 17 00:00:00 2001
From: paulc1983 <paul.cripps@gmail.com>
Date: Wed, 16 Sep 2026 16:27:18 +0100
Subject: [PATCH] Trim whitespace from string values in
 TolerantStringJsonConverter and add unit tests for validation

---
 .../TolerantStringJsonConverter.cs            |   2 +-
 .../TolerantStringJsonConverterTests.cs       | 128 ++++++++++++++++++
 2 files changed, 129 insertions(+), 1 deletion(-)
 create mode 100644 tests/DfE.CheckPerformanceData.UnitTests/ResultsEnquiry/TolerantStringJsonConverterTests.cs

diff --git a/src/DfE.CheckPerformanceData.Application/ResultsEnquiry/TolerantStringJsonConverter.cs b/src/DfE.CheckPerformanceData.Application/ResultsEnquiry/TolerantStringJsonConverter.cs
index 8cfddb2da..f21a738fa 100644
--- a/src/DfE.CheckPerformanceData.Application/ResultsEnquiry/TolerantStringJsonConverter.cs
+++ b/src/DfE.CheckPerformanceData.Application/ResultsEnquiry/TolerantStringJsonConverter.cs
@@ -29,7 +29,7 @@ public override string Read(ref Utf8JsonReader reader, Type typeToConvert, JsonS
         => reader.TokenType switch
         {
             JsonTokenType.Null => string.Empty,
-            JsonTokenType.String => reader.GetString() ?? string.Empty,
+            JsonTokenType.String => (reader.GetString() ?? string.Empty).Trim(),
             // The raw token text, so 5 reads as "5" and not "5.0" via a round-trip through double.
             JsonTokenType.Number => Encoding.UTF8.GetString(
                 reader.HasValueSequence ? BuffersExtensions.ToArray(reader.ValueSequence) : reader.ValueSpan),
diff --git a/tests/DfE.CheckPerformanceData.UnitTests/ResultsEnquiry/TolerantStringJsonConverterTests.cs b/tests/DfE.CheckPerformanceData.UnitTests/ResultsEnquiry/TolerantStringJsonConverterTests.cs
new file mode 100644
index 000000000..66e17e925
--- /dev/null
+++ b/tests/DfE.CheckPerformanceData.UnitTests/ResultsEnquiry/TolerantStringJsonConverterTests.cs
@@ -0,0 +1,128 @@
+using System.Text.Json;
+using System.Text.Json.Serialization;
+using DfE.CheckPerformanceData.Application.ResultsEnquiry;
+using DfE.CheckPerformanceData.Infrastructure.BlobStorage;
+
+namespace DfE.CheckPerformanceData.Application.UnitTests.ResultsEnquiry;
+
+public sealed class TolerantStringJsonConverterTests
+{
+    private static readonly JsonSerializerOptions Opts = ResultsEnquiryJson.Options;
+
+    // Helper: deserializes a JSON object with a single "v" property through the production converter.
+    private static string ReadValue(string json)
+        => JsonSerializer.Deserialize<Wrapper>(json, Opts)!.V;
+
+    [Fact]
+    public void Reads_a_string_value_as_is()
+    {
+        var result = ReadValue("""{ "v": "hello" }""");
+        Assert.Equal("hello", result);
+    }
+
+    [Fact]
+    public void Reads_null_as_empty_string()
+    {
+        var result = ReadValue("""{ "v": null }""");
+        Assert.Equal(string.Empty, result);
+    }
+
+    [Fact]
+    public void Reads_a_number_as_raw_utf8_bytes()
+    {
+        var result = ReadValue("""{ "v": 123 }""");
+        Assert.Equal("123", result);
+    }
+
+    [Fact]
+    public void Reads_true_as_bool_true_string()
+    {
+        var result = ReadValue("""{ "v": true }""");
+        Assert.Equal(bool.TrueString, result);
+    }
+
+    [Fact]
+    public void Reads_false_as_bool_false_string()
+    {
+        var result = ReadValue("""{ "v": false }""");
+        Assert.Equal(bool.FalseString, result);
+    }
+
+    [Fact]
+    public void Throws_on_unexpected_token_type()
+    {
+        // "v" points at a JSON array — TolerantStringJsonConverter does not handle arrays.
+        Assert.Throws<JsonException>(() => ReadValue("""{ "v": [1,2] }"""));
+    }
+
+    [Fact]
+    public void Write_produces_valid_json_string()
+    {
+        using var stream = new MemoryStream();
+        using (var writer = new Utf8JsonWriter(stream))
+        {
+            JsonSerializer.Serialize(writer, "hello world", Opts);
+        }
+
+        var json = System.Text.Encoding.UTF8.GetString(stream.ToArray());
+        Assert.Equal("\"hello world\"", json);
+    }
+
+    // --- FR-005: composite key identity ---
+
+    [Fact]
+    public void Trims_leading_and_trailing_whitespace_from_string_fields()
+    {
+        const string json = """
+        [
+          {
+            "CYPMD_ID": "1",
+            "QAN": "  6037116X  ",
+            "QUAL_NAME": "  GCSE (9-1) Bus. Studs  ",
+            "SYLLABUS": "  1BS0  ",
+            "SESSION": "  S2024  ",
+            "GRADE": "  5  ",
+            "SOURCE": "  16to19_MAIN  "
+          }
+        ]
+        """;
+
+        var records = JsonSerializer.Deserialize<List<StudentResultRecord>>(json, StudentResultsBlobClient.JsonOptions)!;
+        var record = records[0];
+
+        Assert.Equal("6037116X", record.Qan);
+        Assert.Equal("GCSE (9-1) Bus. Studs", record.QualificationName);
+        Assert.Equal("1BS0", record.SyllabusCode);
+        Assert.Equal("S2024", record.Session);
+        Assert.Equal("5", record.Grade);
+        Assert.Equal("16to19_MAIN", record.SourceFile);
+    }
+
+    [Fact]
+    public void CompositeKey_uses_trimmed_values()
+    {
+        const string json = """
+        [
+          {
+            "CYPMD_ID": "1",
+            "QAN": "  6037116X  ",
+            "QUAL_NAME": "  GCSE  ",
+            "SYLLABUS": "  1BS0  ",
+            "SESSION": "  S2024  ",
+            "GRADE": "  5  ",
+            "SOURCE": "  16to19_MAIN  "
+          }
+        ]
+        """;
+
+        var record = JsonSerializer.Deserialize<List<StudentResultRecord>>(json, StudentResultsBlobClient.JsonOptions)![0];
+
+        Assert.Equal("6037116X|S2024|16to19_MAIN", record.CompositeKey);
+    }
+
+    private sealed class Wrapper
+    {
+        [JsonPropertyName("v")]
+        public string V { get; init; } = string.Empty;
+    }
+}