rowGroupInfos = Lists.newArrayList();
long start = 0, length = 0;
rowCount = 0;
columnValueCounts = new HashMap<SchemaPath, Long>();
ColumnChunkMetaData columnChunkMetaData;
for (FileStatus status : statuses) {
List<Footer> footers = ParquetFileReader.readFooters(formatPlugin.getHadoopConfig(), status);
if (footers.size() == 0) {
throw new IOException(String.format("Unable to find footer for file %s", status.getPath().getName()));
}
for (Footer footer : footers) {
int index = 0;
ParquetMetadata metadata = footer.getParquetMetadata();
for (BlockMetaData rowGroup : metadata.getBlocks()) {
long valueCountInGrp = 0;
// need to grab block information from HDFS
columnChunkMetaData = rowGroup.getColumns().iterator().next();
start = columnChunkMetaData.getFirstDataPageOffset();
// this field is not being populated correctly, but the column chunks know their sizes, just summing them for
// now
// end = start + rowGroup.getTotalByteSize();
length = 0;
for (ColumnChunkMetaData col : rowGroup.getColumns()) {