[ORC-24] C++ reader for direct string encodings occasionally skips bytes

• Type: Bug
• Status: Closed
• Priority: Major
• Resolution: Fixed
• Fix Version/s: 1.0.0
• Component/s: C++
• Assignee: Owen O'Malley
• Reporter: Owen O'Malley
• Created: 03/Aug/15 20:21
• Updated: 26/Jan/16 00:51
• Resolved: 05/Aug/15 16:21

Description

The ORC C++ direct string column reader can occasionally skip bytes in the blob stream.

The necessary conditions are:

• The column is a string column and is directly encoded.
• The blob stream for the row batch crosses a compression block boundary.
• There is a null value toward the end of the block boundary.
• The value in the length value of the null value crosses the block boundary, but the length value of the following value does not.