Major Description
When reading some parquet files can get an error like:
java.io.IOException: expecting more rows but reached last block. Read 0 out of 1194236
This happens when looking for a needle thats pretty rare in a large haystack.
The issue here I believe is that the total row count is calculated at
But we pass the blocks we filtered via
org.apache.parquet.filter2.compat.RowGroupFilter.filterRowGroups
to the ParquetFileReader constructor.
However the ParquetFileReader constructor will filter the list of blocks again using
if a block is filtered out by the latter method, and not the former the vectorized reader will believe it should see more rows than it will.
the fix I used locally is pretty straight forward:
for (BlockMetaData block : blocks) { this.totalRowCount += block.getRowCount(); }
goes to
this.totalRowCount = this.reader.getRecordCount();
rdblue do you know if this sounds right? The second filter method in the ParquetFileReader might filter more blocks leading to the count being off?
