Fix #285 (add CsvParser.Feature.FAIL_ON_MISSING_HEADER_COLUMNS)

csv/src/main/java/com/fasterxml/jackson/dataformat/csv/CsvParser.java

@@ -3,6 +3,8 @@
 import java.io.*;
 import java.math.BigDecimal;
 import java.math.BigInteger;
+import java.util.LinkedHashSet;
+import java.util.Set;
 
 import com.fasterxml.jackson.core.*;
 import com.fasterxml.jackson.core.base.ParserMinimalBase;
@@ -106,7 +108,7 @@ public enum Feature
         /**
          * Feature that allows failing (with a {@link CsvReadException}) in cases
          * where number of column values encountered is less than number of columns
-         * declared in active schema ("missing columns").
+         * declared in the active schema ("missing columns").
          *<p>
          * Note that this feature has precedence over {@link #INSERT_NULLS_FOR_MISSING_COLUMNS}
          *<p>
@@ -116,6 +118,17 @@ public enum Feature
          */
         FAIL_ON_MISSING_COLUMNS(false),
 
+        /**
+         * Feature that allows failing (with a {@link CsvReadException}) in cases
+         * where number of header columns encountered is less than number of columns
+         * declared in the active schema (if there is one).
+         *<p>
+         * Feature is enabled by default
+         *
+         * @since 2.14
+         */
+        FAIL_ON_MISSING_HEADER_COLUMNS(true),
+
         /**
          * Feature that allows "inserting" virtual key / `null` value pairs in case
          * a row contains fewer columns than declared by configured schema.
@@ -784,7 +797,8 @@ protected void _readHeaderLine() throws IOException {
                default schema based on the columns found in the header.
          */
 
-        if (_schema.size() > 0 && !_schema.reordersColumns()) {
+        final int schemaColumnCount = _schema.size();
+        if (schemaColumnCount > 0 && !_schema.reordersColumns()) {
             if (_schema.strictHeaders()) {
                 String name;
                 int ix = 0;
@@ -840,13 +854,24 @@ protected void _readHeaderLine() throws IOException {
 
         // Ok: did we get any  columns?
         CsvSchema newSchema = builder.build();
-        int size = newSchema.size();
-        if (size < 2) { // 1 just because we may get 'empty' header name
-            String first = (size == 0) ? "" : newSchema.columnName(0).trim();
+        int newColumnCount = newSchema.size();
+        if (newColumnCount < 2) { // 1 just because we may get 'empty' header name
+            String first = (newColumnCount == 0) ? "" : newSchema.columnName(0).trim();
             if (first.length() == 0) {
                 _reportCsvMappingError("Empty header line: can not bind data");
             }
         }
+        // [dataformats-text#285]: Are we missing something?
+        int diff = schemaColumnCount - newColumnCount;
+        if (diff > 0) {
+            Set<String> oldColumnNames = new LinkedHashSet<>();
+            _schema.getColumnNames(oldColumnNames);
+            oldColumnNames.removeAll(newSchema.getColumnNames());
+            _reportCsvMappingError(String.format("Missing %d header column%s: [\"%s\"]",
+                    diff, (diff == 1) ? "" : "s",
+                            String.join("\",\"", oldColumnNames)));
+        }
+
         // otherwise we will use what we got
         setSchema(builder.build());
     }

csv/src/main/java/com/fasterxml/jackson/dataformat/csv/CsvSchema.java

@@ -927,7 +927,7 @@ public CsvSchema(Column[] columns, int features,
         if (_columns.length == 0) {
             _columnsByName = Collections.emptyMap();
         } else {
-            _columnsByName = new HashMap<String,Column>(4 + _columns.length);
+            _columnsByName = new LinkedHashMap<String,Column>(4 + _columns.length);
             for (Column c : _columns) {
                 _columnsByName.put(c.getName(), c);
             }
@@ -976,7 +976,7 @@ protected CsvSchema(CsvSchema base, Column[] columns)
         if (_columns.length == 0) {
             _columnsByName = Collections.emptyMap();
         } else {
-            _columnsByName = new HashMap<String,Column>(4 + _columns.length);
+            _columnsByName = new LinkedHashMap<String,Column>(4 + _columns.length);
             for (Column c : _columns) {
                 _columnsByName.put(c.getName(), c);
             }
@@ -1407,7 +1407,7 @@ public String getNullValueString() {
     /* Public API, extended; column access
     /**********************************************************************
      */
-    
+
     @Override
     public Iterator<Column> iterator() {
         return Arrays.asList(_columns).iterator();
@@ -1456,7 +1456,31 @@ public Column column(String name, int probableIndex) {
         }
         return _columnsByName.get(name);
     }
-
+
+    /**
+     * Accessor for getting names of included columns, in the order they are
+     * included in the schema.
+     *
+     * @since 2.14
+     */
+    public List<String> getColumnNames() {
+        return (List<String>) getColumnNames(new ArrayList<String>(_columns.length));
+    }
+
+    /**
+     * Accessor for getting names of included columns, added in given
+     * {@code Collection}.
+     *
+     * @since 2.14
+     */
+    public Collection<String> getColumnNames(Collection<String> names) {
+        final int len = _columns.length;
+        for (int i = 0; i < len; ++i) {
+            names.add(_columns[i].getName());
+        }
+        return names;
+    }
+
     /**
      * Method for getting description of column definitions in
      * developer-readable form

csv/src/test/java/com/fasterxml/jackson/dataformat/csv/failing/MissingColumns285Test.java

@@ -28,16 +28,17 @@ public void testMissingWithReorder() throws Exception
                 .addColumn("name").addColumn("age").build();
         final String CSV = "name\n"
                 +"Roger\n";
-        MappingIterator<Map<String, Object>> it = MAPPER
-                .readerFor(Map.class)
-                .with(csvSchema)
-                .readValues(CSV);
+        // Need to have it all inside try block since construction tries to read
+        // the first token
         try {
+            MappingIterator<Map<String, Object>> it = MAPPER
+                    .readerFor(Map.class)
+                    .with(csvSchema)
+                    .readValues(CSV);
             it.nextValue();
             fail("Should not pass with missing columns");
         } catch (CsvReadException e) {
-            verifyException(e, "Not enough column values");
-            verifyException(e, "expected 2, found 1");
+            verifyException(e, "Missing 1 header column: [\"age\"]");
         }
     }
 }

release-notes/CREDITS-2.x

@@ -144,8 +144,11 @@ Francesco Tumanischvili (frantuma@github)
  (2.11.1)
 
 Björn Michael (bjmi@github)
-* Reported #204: `CsvParser.Feature.ALLOW_TRAILING_COMMA` doesn't work with header columns
+* Reported #204: (csv) `CsvParser.Feature.ALLOW_TRAILING_COMMA` doesn't work with header columns
  (2.11.2)
+* Reported #285: (csv) Missing columns from header line (compare to `CsvSchema`) not detected
+  when reordering columns (add `CsvParser.Feature.FAIL_ON_MISSING_HEADER_COLUMNS`)
+ (2.14.0)
 
 Jesper Nielsen (jn-asseco@github)
 * Requested #175: (yaml) Add `YAMLGenerator.Feature.INDENT_ARRAYS_WITH_INDICATOR`

release-notes/VERSION-2.x

@@ -16,6 +16,9 @@ Active Maintainers:
 
 2.14.0 (not yet released)
 
+#285: (csv) Missing columns from header line (compare to `CsvSchema`) not detected
+  when reordering columns (add `CsvParser.Feature.FAIL_ON_MISSING_HEADER_COLUMNS`)
+ (reported by Björn M)
 #297: (csv) CSV schema caching POJOs with different views
  (contributed by Falk H)
 #314: (csv) Add fast floating-point parsing, generation support