[ https://issues.apache.org/jira/browse/HIVE-4844?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13766216#comment-13766216 ]
Phabricator commented on HIVE-4844: ----------------------------------- jdere has commented on the revision "HIVE-4844 [jira] Add varchar data type". Most of the SerDe support has been broken out into HIVE-5161. The SerDe's I'm testing there are RegexSerDe, LazyBinarySerDe, LazySimpleSerDe, ColumnarSerDe, LazyBinaryColumnarSerDe, OrcSerDe. I'll need to take a look again, but if HBase/Avro SerDe's use reflection-based methods to get their Types/ObjectInspectors, then they won't be able to support varchars properly. INLINE COMMENTS ql/src/java/org/apache/hadoop/hive/ql/exec/FunctionRegistry.java:629 will change in the next patch ql/src/java/org/apache/hadoop/hive/ql/parse/DDLSemanticAnalyzer.java:160 will change in next patch ql/src/java/org/apache/hadoop/hive/ql/parse/DDLSemanticAnalyzer.java:169 will change in next patch ql/src/java/org/apache/hadoop/hive/ql/udf/GenericUDFEncode.java:48 sure, will change ql/src/java/org/apache/hadoop/hive/ql/udf/GenericUDFEncode.java:56 will change this message too ql/src/java/org/apache/hadoop/hive/ql/udf/generic/GenericUDFToVarchar.java:36 whoops forgot to add that .. will do ql/src/java/org/apache/hadoop/hive/ql/udf/generic/GenericUDFToVarchar.java:39-40 Actually, my latest patch did mark these fields as transient .. I'll need to update this review with the updated patch. serde/src/java/org/apache/hadoop/hive/serde2/lazy/objectinspector/primitive/LazyHiveVarcharObjectInspector.java:34 Do we need to serialize object inspectors? my impression is that should always have been transient, at least when they're used in UDFs since they can be set during initialize(). serde/src/java/org/apache/hadoop/hive/serde2/objectinspector/primitive/PrimitiveObjectInspectorConverter.java:401 Yeah, this comment is gone in the updated patch. serde/src/java/org/apache/hadoop/hive/serde2/objectinspector/primitive/PrimitiveObjectInspectorConverter.java:474 2011 SQL Standard (6.13 cast specification) does say that the value should be "TRUE"/"FALSE". This also matches what we're doing with the string type. serde/src/java/org/apache/hadoop/hive/serde2/objectinspector/primitive/WritableConstantHiveVarcharObjectInspector.java:33 ok, will change serde/src/java/org/apache/hadoop/hive/serde2/objectinspector/primitive/WritableHiveVarcharObjectInspector.java:35 will change REVISION DETAIL https://reviews.facebook.net/D12699 To: JIRA, jdere Cc: ashutoshc > Add varchar data type > --------------------- > > Key: HIVE-4844 > URL: https://issues.apache.org/jira/browse/HIVE-4844 > Project: Hive > Issue Type: New Feature > Components: Types > Reporter: Jason Dere > Assignee: Jason Dere > Attachments: HIVE-4844.10.patch, HIVE-4844.11.patch, > HIVE-4844.12.patch, HIVE-4844.13.patch, HIVE-4844.14.patch, > HIVE-4844.15.patch, HIVE-4844.16.patch, HIVE-4844.1.patch.hack, > HIVE-4844.2.patch, HIVE-4844.3.patch, HIVE-4844.4.patch, HIVE-4844.5.patch, > HIVE-4844.6.patch, HIVE-4844.7.patch, HIVE-4844.8.patch, HIVE-4844.9.patch, > HIVE-4844.D12699.1.patch, screenshot.png > > > Add new varchar data types which have support for more SQL-compliant > behavior, such as SQL string comparison semantics, max length, etc. > Char type will be added as another task. -- This message is automatically generated by JIRA. If you think it was sent incorrectly, please contact your JIRA administrators For more information on JIRA, see: http://www.atlassian.com/software/jira