Merge pull request #59 from NetTopologySuite/fix/issue046

Prevent duplicate field names. Fix #46.

Author: KubaSzostak

src/Directory.Build.props

@@ -99,7 +99,7 @@
 
   <ItemGroup>
     <!-- SourceLink adds stuff to let debuggers step into our code. -->
-    <PackageReference Include="Microsoft.SourceLink.GitHub" Version="1.0.0" PrivateAssets="All" />
+    <PackageReference Include="Microsoft.SourceLink.GitHub" Version="8.0.0" PrivateAssets="All" />
   </ItemGroup>
 
 </Project>

src/NetTopologySuite.IO.Esri.Shapefile/Dbf/DbfReader.cs

@@ -121,10 +121,15 @@ private void Initialize(Stream stream, Encoding encoding = null)
 
             Buffer.AssignFrom(DbfStream, fieldsHeaderSize);
             Fields = new DbfFieldCollection(fieldCount);
+            var duplicates = new Dictionary<string, int>();
             for (int i = 0; i < fieldCount; i++)
             {
                 var field = Buffer.ReadDbaseFieldDescriptor(Encoding);
+                string originalName = field.Name;
+                duplicates.TryGetValue(originalName, out int duplicateCount);
+                field.DuplicateCount = duplicateCount;
                 Fields.Add(field);
+                duplicates[originalName] = ++duplicateCount;
                 //Binary.TraceToConsole("Field Header: " + Fields[Fields.Count -1].Name, i * Dbf.FieldDescriptorSize, Dbf.FieldDescriptorSize);
             }
 

src/NetTopologySuite.IO.Esri.Shapefile/Dbf/DbfStreamExtensions.cs

@@ -84,8 +84,7 @@ public static Encoding ReadDbfEncoding(this Stream stream)
         public static void WriteDbaseFieldDescriptor(this Stream stream, DbfField field, Encoding encoding)
         {
             encoding = encoding ?? Dbf.DefaultEncoding;
-            var name = field.Name.PadRight(Dbf.MaxFieldNameLength, char.MinValue); // Field name must have empty space zero-filled 
-
+            var name = field.OriginalName.PadRight(Dbf.MaxFieldNameLength, char.MinValue); // Field name must have empty space zero-filled 
 
             stream.WriteString(name, Dbf.MaxFieldNameLength, encoding);
             stream.WriteNullBytes(1);

src/NetTopologySuite.IO.Esri.Shapefile/Dbf/Fields/DbfField.cs

@@ -15,10 +15,22 @@ public abstract class DbfField
     {
         // TODO: Spit DbfField into DbfFieldDefinition (Without Value property) and DbfFieldValue
 
+        /// <summary>
+        /// Original field name
+        /// </summary>
+        internal string OriginalName { get; }
+
         /// <summary>
         /// Field Name.
         /// </summary>
-        public string Name { get; }
+        public string Name
+        {
+            get
+            {
+                string suffix = DuplicateCount > 0 ? $"_{DuplicateCount}" : string.Empty;
+                return $"{OriginalName}{suffix}";
+            }
+        }
 
         /// <summary>
         /// Field Type.
@@ -75,17 +87,12 @@ internal DbfField(string name, DbfType type, int length, int precision)
                 throw new ArgumentException($"Ivalid dBASE field precision: {precision}.", nameof(precision));
             }
 
-            Name = name;
+            OriginalName = name;
             FieldType = type;
             Length = length;
             NumericScale = precision;
         }
 
-        private bool IsValidFieldNameChar(char c)
-        {
-            return (c >= 'a' && c <= 'z') || (c >= 'A' && c <= 'Z') || c == '_' || (c >= '0' && c <= '9');
-        }
-
         /// <inheritdoc/>
         public override string ToString()
         {
@@ -113,7 +120,7 @@ internal Exception GetFieldValueError(object value, string additionalDescription
         /// </summary>
         public virtual bool IsNull => Value == null;
 
-
+        internal int DuplicateCount { get; set; }
 
         internal static readonly IAttributesTable EmptyFieldValues = new AttributesTable();
 

src/NetTopologySuite.IO.Esri.Shapefile/Dbf/Fields/DbfFieldCollection.cs

@@ -59,6 +59,8 @@ IEnumerator IEnumerable.GetEnumerator()
 
         internal void Add(DbfField field)
         {
+            if (FieldDictionary.ContainsKey(field.Name))
+                throw new ArgumentException($"Field with name '{field.Name}' already defined.");
             Fields.Add(field);
             FieldDictionary.Add(field.Name, field);
         }

test/Directory.Build.props

@@ -5,8 +5,8 @@
 
   <ItemGroup>
     <PackageReference Include="nunit" Version="3.12.0" />
-    <PackageReference Include="NUnit3TestAdapter" Version="3.16.1" />
-    <PackageReference Include="Microsoft.NET.Test.Sdk" Version="16.5.0" />
+    <PackageReference Include="NUnit3TestAdapter" Version="4.6.0" />
+    <PackageReference Include="Microsoft.NET.Test.Sdk" Version="17.11.1" />
   </ItemGroup>
 
 </Project>

test/NetTopologySuite.IO.Esri.Test/Issues/Issue046.cs

@@ -0,0 +1,22 @@
+using NetTopologySuite.IO.Esri.Dbf;
+using NUnit.Framework;
+using System.Collections.Generic;
+
+namespace NetTopologySuite.IO.Esri.Test.Issues
+{
+    internal class Issue046
+    {
+        [Test]
+        public void TestNoDuplicateColumnName()
+        {
+            using var dbfReader = new DbfReader(TestShapefiles.PathTo("Issues/046/Antragsschläge 2024_POLYGONE.dbf"));
+            var dbfFields = dbfReader.Fields;
+            var fieldNames = new HashSet<string>();
+            for (int i = 0; i < dbfFields.Count; i++)
+            {
+                Assert.That(fieldNames.Contains(dbfFields[i].Name), Is.False);
+                fieldNames.Add(dbfFields[i].Name);
+            }
+        }
+    }
+}