From 163f22757444b77540902be442aa1a7befa8ee82 Mon Sep 17 00:00:00 2001
From: Ryan Davis <rdavis@infiniteenergy.com>
Date: Tue, 27 Apr 2021 14:41:49 -0400
Subject: [PATCH] treat missing rows as a mismatch

Throwing `KeyNotFound` is not helpful for larger analsyis.

Had to loosed a lot of non-null strings because we won't have any context from the "right" file, so need to pass some more nulls around when building output.
---
 src/XlsxCompare.Tests/XlsxComparerTests.cs | 50 ++++++++++++++++++++++
 src/XlsxCompare/Mismatch.cs                |  2 +-
 src/XlsxCompare/ResultsWriter.cs           |  6 +--
 src/XlsxCompare/XlsxComparer.cs            | 12 +++++-
 4 files changed, 64 insertions(+), 6 deletions(-)

diff --git a/src/XlsxCompare.Tests/XlsxComparerTests.cs b/src/XlsxCompare.Tests/XlsxComparerTests.cs
index 3d3c076..1ee2e1e 100644
--- a/src/XlsxCompare.Tests/XlsxComparerTests.cs
+++ b/src/XlsxCompare.Tests/XlsxComparerTests.cs
@@ -61,5 +61,55 @@ public void Compare_SomeErrors_ReturnsCorrectResult()
                 Assert.AreEqual(expectedContext[id], newId);
             }
         }
+
+        [TestMethod]
+        public void Compare_NoJoinMatchesWithContext_ReturnsMissingDataAsMismatch()
+        {
+            var opts = new CompareOptions(
+                LeftKeyColumn: "id",
+                RightKeyColumn: "new_id",
+                new[]{
+                    new Assertion("Email", "EML")
+                },
+                new ResultOptions(
+                    LeftColumnNames: new[] { "Name" }
+                )
+            );
+            var results = _comparer.Compare("left.xlsx", "right.xlsx", opts);
+
+            Assert.AreEqual(3, results.TotalCellMismatches);
+            Assert.AreEqual(3, results.TotalRowMismatches);
+
+            var expectedContext = new Dictionary<string, string>(){
+                {"1", "leading trailing spaces"},
+                {"2", "null email"},
+                {"3", "whitespace address"}
+            };
+
+            foreach (var mismatch in results.Mismatches)
+            {
+                var id = mismatch.Key;
+                var name = mismatch.Context["Name"];
+                Assert.AreEqual(expectedContext[id], name);
+            }
+        }
+
+        [TestMethod]
+        public void Compare_NoJoinMatchesWithIgnoringMissingRows_ReturnsNoMismatches()
+        {
+            var opts = new CompareOptions(
+                LeftKeyColumn: "id",
+                RightKeyColumn: "new_id",
+                Assertions: new[]{
+                    new Assertion("Email", "EML")
+                },
+                ResultOptions: new ResultOptions(),
+                IgnoreMissingRows: true
+            );
+            var results = _comparer.Compare("left.xlsx", "right.xlsx", opts);
+
+            Assert.AreEqual(0, results.TotalCellMismatches);
+            Assert.AreEqual(0, results.TotalRowMismatches);
+        }
     }
 }
diff --git a/src/XlsxCompare/Mismatch.cs b/src/XlsxCompare/Mismatch.cs
index bef27e8..fee887d 100644
--- a/src/XlsxCompare/Mismatch.cs
+++ b/src/XlsxCompare/Mismatch.cs
@@ -6,7 +6,7 @@ record Mismatch(
         Assertion Assertion,
         string Key,
         string LeftValue,
-        string RightValue,
+        string? RightValue,
         IReadOnlyDictionary<string, string> Context
     );
 }
diff --git a/src/XlsxCompare/ResultsWriter.cs b/src/XlsxCompare/ResultsWriter.cs
index a11a595..024d8ef 100644
--- a/src/XlsxCompare/ResultsWriter.cs
+++ b/src/XlsxCompare/ResultsWriter.cs
@@ -57,11 +57,11 @@ private static IEnumerable<string> GetHeaders(ResultOptions opts)
             yield return opts.RightValueHeader;
         }
 
-        private static IEnumerable<string> GetValues(ResultOptions opts, Mismatch mismatch)
+        private static IEnumerable<string?> GetValues(ResultOptions opts, Mismatch mismatch)
         {
             foreach (var col in opts.ContextColumnNames)
             {
-                yield return mismatch.Context[col];
+                yield return mismatch.Context.GetValueOrDefault(col);
             }
             yield return mismatch.Assertion.LeftColumnName;
             // TODO: maybe format?
@@ -69,7 +69,7 @@ private static IEnumerable<string> GetValues(ResultOptions opts, Mismatch mismat
             yield return mismatch.RightValue;
         }
 
-        private static void WriteRow(ExcelWorksheet sheet, int row, IEnumerable<string> values)
+        private static void WriteRow(ExcelWorksheet sheet, int row, IEnumerable<string?> values)
         {
             var col = 1;
             foreach (var value in values)
diff --git a/src/XlsxCompare/XlsxComparer.cs b/src/XlsxCompare/XlsxComparer.cs
index 1e4acf0..a270322 100644
--- a/src/XlsxCompare/XlsxComparer.cs
+++ b/src/XlsxCompare/XlsxComparer.cs
@@ -29,13 +29,22 @@ public CompareResult Compare(string leftPath, string rightPath, CompareOptions o
         private IEnumerable<Mismatch> Compare(XlsxFacade left, XlsxFacade right, CompareOptions opts)
         {
             _logger.LogInformation("Comparing {LeftXlsx} to {RightXlsx}", left, right);
+            var joinAssertion = new Assertion(opts.LeftKeyColumn, opts.RightKeyColumn);
             foreach (var leftRow in left.Rows)
             {
                 var key = left.GetSafeValue(leftRow, opts.LeftKeyColumn);
+                var leftContext = new Dictionary<string, string>(GetContext(left, leftRow, opts.ResultOptions.LeftColumnNames));
                 if (!right.TryFindRow(opts.RightKeyColumn, key, out var rightRow))
                 {
                     if (opts.IgnoreMissingRows) { continue; }
-                    throw new KeyNotFoundException($"Could not find '{key}' in {opts.RightKeyColumn}");
+                    yield return new Mismatch(
+                        Assertion: joinAssertion,
+                        Key: key,
+                        LeftValue: key,
+                        RightValue: null,
+                        Context: leftContext
+                    );
+                    continue;
                 }
 
                 foreach (var assertion in opts.Assertions)
@@ -45,7 +54,6 @@ private IEnumerable<Mismatch> Compare(XlsxFacade left, XlsxFacade right, Compare
 
                     if (!assertion.IsMatch(leftValue, rightValue))
                     {
-                        var leftContext = GetContext(left, leftRow, opts.ResultOptions.LeftColumnNames);
                         var rightContext = GetContext(right, rightRow, opts.ResultOptions.RightColumnNames);
                         var context = new Dictionary<string, string>(
                             leftContext.Concat(rightContext));