PARQUET-511: Integer overflow when counting values in column.

This commit fixes an issue when the number of entries in a column page is larger than the size of an integer. No exception is thrown directly, but the def level is set incorrectly, leading to a null value being returned during read.

Author: goreckm

parquet-column/src/main/java/org/apache/parquet/column/impl/ColumnReaderImpl.java

@@ -150,7 +150,7 @@ public double getDouble() {
   private int dictionaryId;
 
   private long endOfPageValueCount;
-  private int readValues = 0;
+  private long readValues = 0;
   private int pageValueCount = 0;
 
   private final PrimitiveConverter converter;
@@ -352,7 +352,7 @@ public ColumnReaderImpl(ColumnDescriptor path, PageReader pageReader, PrimitiveC
       this.dictionary = null;
     }
     this.totalValueCount = pageReader.getTotalValueCount();
-    if (totalValueCount == 0) {
+    if (totalValueCount <= 0) {
       throw new ParquetDecodingException("totalValueCount == 0");
     }
     consume();

parquet-hadoop/src/main/java/org/apache/parquet/hadoop/ColumnChunkPageReadStore.java

@@ -63,7 +63,7 @@ static final class ColumnChunkPageReader implements PageReader {
       this.decompressor = decompressor;
       this.compressedPages = new LinkedList<DataPage>(compressedPages);
       this.compressedDictionaryPage = compressedDictionaryPage;
-      int count = 0;
+      long count = 0;
       for (DataPage p : compressedPages) {
         count += p.getValueCount();
       }