[HIVE-3253] ArrayIndexOutOfBounds exception for deeply nested structs

• Type: Bug
• Status: Closed
• Priority: Major
• Resolution: Fixed
• Affects Version/s: 0.9.0, 0.10.0
• Fix Version/s: 0.12.0
• Component/s: Serializers/Deserializers

Release Note

This change increases the number of levels of nesting supported in hive select queries. The limitation in the serialization format used by File Output Operator for these queries (LazySimpleSerde) was restricted the number of levels of nesting to 8 earlier, this has now been extended to 24. This extended levels of nesting is turned on by default.

This change also improves the number of levels of nesting that you can use with tables that use LazySimpleSerde. It uses additional control charactors as delimiters. This means that your data should not have these charactors or you need to escape these charactors. As this change introduces a new requirement for the way data has been written, this is not backward compatible. Hence this is not enabled by default. To enabled this, you need to set the serde property hive.serialization.extend.nesting.levels to true.

Look at 'ESCAPED BY' documentation for create-table, to learn how to enable escaping of the delimiter charactors. https://cwiki.apache.org/confluence/display/Hive/LanguageManual+DDL

• Assignee: Thejas Nair
• Reporter: Swarnim Kulkarni
• Created: 12/Jul/12 13:26
• Updated: 28/Jan/15 18:11
• Resolved: 03/Jul/13 15:53

Description

It was observed that while creating table with deeply nested structs might throw this exception:

java.lang.ArrayIndexOutOfBoundsException: 9
        at org.apache.hadoop.hive.serde2.lazy.LazyFactory.createLazyObjectInspector(LazyFactory.java:281)
	at org.apache.hadoop.hive.serde2.lazy.LazyFactory.createLazyObjectInspector(LazyFactory.java:263)
	at org.apache.hadoop.hive.serde2.lazy.LazyFactory.createLazyObjectInspector(LazyFactory.java:276)
	at org.apache.hadoop.hive.serde2.lazy.LazyFactory.createLazyObjectInspector(LazyFactory.java:263)
	at org.apache.hadoop.hive.serde2.lazy.LazyFactory.createLazyObjectInspector(LazyFactory.java:276)
	at org.apache.hadoop.hive.serde2.lazy.LazyFactory.createLazyObjectInspector(LazyFactory.java:263)
	at org.apache.hadoop.hive.serde2.lazy.LazyFactory.createLazyObjectInspector(LazyFactory.java:276)
	at org.apache.hadoop.hive.serde2.lazy.LazyFactory.createLazyObjectInspector(LazyFactory.java:263)
	at org.apache.hadoop.hive.serde2.lazy.LazyFactory.createLazyObjectInspector(LazyFactory.java:276)
	at org.apache.hadoop.hive.serde2.lazy.LazyFactory.createLazyStructInspector(LazyFactory.java:354)

The reason being that currently the separators array has been hardcoded to be of size 8 in the LazySimpleSerde.

// Read the separators: We use 8 levels of separators by default, but we
// should change this when we allow users to specify more than 10 levels
// of separators through DDL.
serdeParams.separators = new byte[8];

If possible, we should increase this size or at least make it configurable to properly handle deeply nested structs.

Issue Links

• is duplicated by: HIVE-4571 Reinvestigate HIVE-337 induced limit on number of separator characters in LazySerDe
• is related to: HIVE-9500 Support nested structs over 24 levels.