fix(sonar): size hash-based collections with the JDK 19+ factory methods (S6485)

new HashSet<>(n)/new HashMap<>(n)/new LinkedHashMap<>(n) treat n as raw
initial capacity, not expected entry count -- undersized relative to
the 0.75 load factor, so these were one resize away from their stated
bound in the worst case (the manual "+1" fudges in DictColumnState show
the sizing was already an approximation). HashSet.newHashSet(n) /
HashMap.newHashMap(n) / LinkedHashMap.newLinkedHashMap(n) take the
actual expected element count and size correctly, guaranteed no resize.

Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01KW6yskY3pE759XUg2UUtf8

Author: dfa1

reader/src/main/java/io/github/dfa1/vortex/reader/ScanIterator.java

@@ -255,7 +255,7 @@ private static SequencedMap<ColumnName, Chunk.Column> expandStruct(StructArray s
         List<ColumnName> names = sd.fieldNames();
         List<DType> types = sd.fieldTypes();
         int n = names.size();
-        var map = new LinkedHashMap<ColumnName, Chunk.Column>(n);
+        var map = LinkedHashMap.<ColumnName, Chunk.Column>newLinkedHashMap(n);
         for (int i = 0; i < n; i++) {
             map.put(names.get(i), new Chunk.Column(sa.field(i), types.get(i)));
         }
@@ -266,7 +266,7 @@ private static SequencedMap<ColumnName, Chunk.Column> expandStruct(StructArray s
 
     private static SequencedMap<ColumnName, Chunk.Column> limitedColumns(
             SequencedMap<ColumnName, Chunk.Column> columns, long rows) {
-        var result = new LinkedHashMap<ColumnName, Chunk.Column>(columns.size());
+        var result = LinkedHashMap.<ColumnName, Chunk.Column>newLinkedHashMap(columns.size());
         for (var entry : columns.entrySet()) {
             Chunk.Column col = entry.getValue();
             result.put(entry.getKey(), new Chunk.Column(Array.limited(col.array(), rows), col.dtype()));
@@ -669,7 +669,7 @@ private SequencedMap<ColumnName, Chunk.Column> buildColumnMap(ChunkSpec chunk, A
             }
             return singleColumn(arr);
         }
-        var scratch = new LinkedHashMap<ColumnName, Chunk.Column>(n);
+        var scratch = LinkedHashMap.<ColumnName, Chunk.Column>newLinkedHashMap(n);
         for (int i = 0; i < n; i++) {
             scratch.put(projectedNames.get(i), new Chunk.Column(
                     decodeOrSlice(i, layouts[i], sliceOffsets[i], chunk.rowCount(), arena),
@@ -692,7 +692,7 @@ private SequencedMap<ColumnName, Chunk.Column> buildSelfContainedColumnMap(Chunk
             }
             return singleColumn(arr);
         }
-        var scratch = new LinkedHashMap<ColumnName, Chunk.Column>(n);
+        var scratch = LinkedHashMap.<ColumnName, Chunk.Column>newLinkedHashMap(n);
         for (int i = 0; i < n; i++) {
             scratch.put(projectedNames.get(i), new Chunk.Column(
                     decodeOrSliceSelfContained(i, layouts[i], sliceOffsets[i], chunk.rowCount(), arena),
@@ -702,7 +702,7 @@ private SequencedMap<ColumnName, Chunk.Column> buildSelfContainedColumnMap(Chunk
     }
 
     private SequencedMap<ColumnName, Chunk.Column> singleColumn(Array array) {
-        var map = new LinkedHashMap<ColumnName, Chunk.Column>(1);
+        var map = LinkedHashMap.<ColumnName, Chunk.Column>newLinkedHashMap(1);
         map.put(projectedNames.getFirst(), new Chunk.Column(array, projectedDtypes.getFirst()));
         return unmodifiable(map);
     }

writer/src/main/java/io/github/dfa1/vortex/writer/DictColumnState.java

@@ -147,7 +147,7 @@ boolean ingestDictChunk(Object data) {
         // First pass: would this chunk's fresh distinct values push the map past the cap? Count them
         // without mutating so the whole chunk is either ingested or rejected atomically — a partial
         // ingest would corrupt the dictionary when the caller demotes on rejection.
-        var pendingNew = new HashSet<>(Math.min(cap, len) + 1);
+        var pendingNew = HashSet.newHashSet(Math.min(cap, len));
         for (int i = 0; i < len; i++) {
             if (validity != null && !validity[i]) {
                 continue;
@@ -387,7 +387,7 @@ static boolean isUtf8DictCandidate(String[] data, boolean[] validity) {
         if (data.length == 0) {
             return false;
         }
-        var seen = new java.util.HashSet<String>(Math.min(GLOBAL_DICT_MAX_CARDINALITY_UTF8, data.length) + 1);
+        var seen = HashSet.<String>newHashSet(Math.min(GLOBAL_DICT_MAX_CARDINALITY_UTF8, data.length));
         for (int i = 0; i < data.length; i++) {
             if ((validity != null && !validity[i]) || data[i] == null) {
                 continue;
@@ -438,7 +438,7 @@ static boolean isDictCandidate(PType ptype, Object data, boolean[] validity) {
         if (n == 0) {
             return false;
         }
-        var seen = new HashSet<>(GLOBAL_DICT_MAX_CARDINALITY + 1);
+        var seen = HashSet.newHashSet(GLOBAL_DICT_MAX_CARDINALITY);
         for (int i = 0; i < n; i++) {
             if (validity != null && !validity[i]) {
                 continue;

writer/src/main/java/io/github/dfa1/vortex/writer/encode/ArrayStats.java

@@ -44,7 +44,7 @@ public static ArrayStats compute(PType ptype, Object data, StatsOptions options)
             return new ArrayStats(n, -1, 0, 0);
         }
         HashMap<Long, int[]> counts = options.countDistinct() || options.trackMostFrequent()
-                                              ? new HashMap<>(Math.min(n, 1 << 16))
+                                              ? HashMap.newHashMap(Math.min(n, 1 << 16))
                                               : null;
         long topFreqBits = 0;
         int topFreq = 0;