[HIVE-20827] Inconsistent results for empty arrays

• Type: Bug
• Status: Closed
• Priority: Major
• Resolution: Fixed
• Fix Version/s: 4.0.0-alpha-1
• Labels: pull-request-available
• Assignee: Teddy Choi
• Reporter: Teddy Choi
• Created: 29/Oct/18 09:31
• Updated: 17/Nov/22 08:54
• Resolved: 06/Dec/18 15:03

Description

LazySimpleDeserializeRead parses an empty array wrong. For example, a line ',' in a text file table with a delimiter ',' and schema 'array<int>, array<array<string>>' shows [null], [[""]], instead of [], [] with MapReduce engine and vectorized execution enabled. LazySimpleDeserializeRead has following code; 

switch (complexField.complexCategory) {
case LIST:
  {
    // Allow for empty string, etc.
    final boolean isNext = (fieldPosition <= complexFieldEnd);

Empty string value read should be only applied to string families, not to other data types. 

Issue Links

• links to: GitHub Pull Request #480