You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@hive.apache.org by Barna Zsombor Klara <zs...@cloudera.com> on 2017/05/22 11:59:23 UTC

Review Request 59446: HIVE-16559: Parquet schema evolution for partitioned tables may break if table and partition serdes differ

-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/59446/
-----------------------------------------------------------

Review request for hive and Sergio Pena.


Bugs: HIVE-16559
    https://issues.apache.org/jira/browse/HIVE-16559


Repository: hive-git


Description
-------

HIVE-16559: Parquet schema evolution for partitioned tables may break if table and partition serdes differ


Diffs
-----

  ql/src/test/queries/clientpositive/parquet_schema_evolution.q e767b8186ead8e005ef4ee40f583a8afc48fa175 
  ql/src/test/results/clientpositive/parquet_schema_evolution.q.out 43d75dc105ff104946d093d7c5e3c9de1cdcb196 
  serde/src/java/org/apache/hadoop/hive/serde2/AbstractSerDe.java a2a85b3dc44d7d7d676ad17eb476edbb5560f1cc 
  serde/src/java/org/apache/hadoop/hive/serde2/SerDeUtils.java b9d23dd72580a684788747bfdff4c35b08726fba 
  serde/src/test/org/apache/hadoop/hive/serde2/TestSerDeUtils.java PRE-CREATION 


Diff: https://reviews.apache.org/r/59446/diff/1/


Testing
-------

Added unit tests and a qtest. Manually tested that no regression is caused for avro and textfile SerDes when columns are added or replaced in a partitioned table.


Thanks,

Barna Zsombor Klara


Re: Review Request 59446: HIVE-16559: Parquet schema evolution for partitioned tables may break if table and partition serdes differ

Posted by Sergio Pena <se...@cloudera.com>.
-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/59446/#review178921
-----------------------------------------------------------


Ship it!




Ship It!

- Sergio Pena


On June 26, 2017, 11:48 a.m., Barna Zsombor Klara wrote:
> 
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/59446/
> -----------------------------------------------------------
> 
> (Updated June 26, 2017, 11:48 a.m.)
> 
> 
> Review request for hive and Sergio Pena.
> 
> 
> Bugs: HIVE-16559
>     https://issues.apache.org/jira/browse/HIVE-16559
> 
> 
> Repository: hive-git
> 
> 
> Description
> -------
> 
> HIVE-16559: Parquet schema evolution for partitioned tables may break if table and partition serdes differ
> 
> 
> Diffs
> -----
> 
>   ql/src/java/org/apache/hadoop/hive/ql/exec/DDLTask.java 87928ee930b5ee974d5e4144a584773a243f8d6f 
>   ql/src/test/queries/clientnegative/parquet_alter_part_table_drop_columns.q PRE-CREATION 
>   ql/src/test/results/clientnegative/parquet_alter_part_table_drop_columns.q.out PRE-CREATION 
> 
> 
> Diff: https://reviews.apache.org/r/59446/diff/3/
> 
> 
> Testing
> -------
> 
> Added a negative qtest. Manually tested that no regression is caused for avro and textfile SerDes when columns are added or replaced in a partitioned table.
> 
> 
> Thanks,
> 
> Barna Zsombor Klara
> 
>


Re: Review Request 59446: HIVE-16559: Parquet schema evolution for partitioned tables may break if table and partition serdes differ

Posted by Barna Zsombor Klara <zs...@cloudera.com>.
-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/59446/
-----------------------------------------------------------

(Updated June 26, 2017, 11:48 a.m.)


Review request for hive and Sergio Pena.


Changes
-------

Removed the new error message and added a warning to clarify the error before the exception.


Bugs: HIVE-16559
    https://issues.apache.org/jira/browse/HIVE-16559


Repository: hive-git


Description
-------

HIVE-16559: Parquet schema evolution for partitioned tables may break if table and partition serdes differ


Diffs (updated)
-----

  ql/src/java/org/apache/hadoop/hive/ql/exec/DDLTask.java 87928ee930b5ee974d5e4144a584773a243f8d6f 
  ql/src/test/queries/clientnegative/parquet_alter_part_table_drop_columns.q PRE-CREATION 
  ql/src/test/results/clientnegative/parquet_alter_part_table_drop_columns.q.out PRE-CREATION 


Diff: https://reviews.apache.org/r/59446/diff/3/

Changes: https://reviews.apache.org/r/59446/diff/2-3/


Testing
-------

Added a negative qtest. Manually tested that no regression is caused for avro and textfile SerDes when columns are added or replaced in a partitioned table.


Thanks,

Barna Zsombor Klara


Re: Review Request 59446: HIVE-16559: Parquet schema evolution for partitioned tables may break if table and partition serdes differ

Posted by Sergio Pena <se...@cloudera.com>.
-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/59446/#review178683
-----------------------------------------------------------




ql/src/java/org/apache/hadoop/hive/ql/ErrorMsg.java
Lines 463 (patched)
<https://reviews.apache.org/r/59446/#comment252832>

    What's the difference with REPLACE_CANNOT_DROP_COLUMNS ? Seems they are use for the same error message.



ql/src/java/org/apache/hadoop/hive/ql/ErrorMsg.java
Lines 464 (patched)
<https://reviews.apache.org/r/59446/#comment252831>

    Should we use CASCADE in capital case for better info that it is a keyword?


- Sergio Pena


On June 19, 2017, 9:52 a.m., Barna Zsombor Klara wrote:
> 
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/59446/
> -----------------------------------------------------------
> 
> (Updated June 19, 2017, 9:52 a.m.)
> 
> 
> Review request for hive and Sergio Pena.
> 
> 
> Bugs: HIVE-16559
>     https://issues.apache.org/jira/browse/HIVE-16559
> 
> 
> Repository: hive-git
> 
> 
> Description
> -------
> 
> HIVE-16559: Parquet schema evolution for partitioned tables may break if table and partition serdes differ
> 
> 
> Diffs
> -----
> 
>   ql/src/java/org/apache/hadoop/hive/ql/ErrorMsg.java 6651900e79a5c3d4ad8329afbe3894544ce9f46e 
>   ql/src/java/org/apache/hadoop/hive/ql/exec/DDLTask.java 87928ee930b5ee974d5e4144a584773a243f8d6f 
>   ql/src/test/queries/clientnegative/parquet_alter_part_table_drop_columns.q PRE-CREATION 
>   ql/src/test/results/clientnegative/parquet_alter_part_table_drop_columns.q.out PRE-CREATION 
> 
> 
> Diff: https://reviews.apache.org/r/59446/diff/2/
> 
> 
> Testing
> -------
> 
> Added a negative qtest. Manually tested that no regression is caused for avro and textfile SerDes when columns are added or replaced in a partitioned table.
> 
> 
> Thanks,
> 
> Barna Zsombor Klara
> 
>


Re: Review Request 59446: HIVE-16559: Parquet schema evolution for partitioned tables may break if table and partition serdes differ

Posted by Barna Zsombor Klara <zs...@cloudera.com>.
-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/59446/
-----------------------------------------------------------

(Updated June 19, 2017, 9:52 a.m.)


Review request for hive and Sergio Pena.


Bugs: HIVE-16559
    https://issues.apache.org/jira/browse/HIVE-16559


Repository: hive-git


Description
-------

HIVE-16559: Parquet schema evolution for partitioned tables may break if table and partition serdes differ


Diffs (updated)
-----

  ql/src/java/org/apache/hadoop/hive/ql/ErrorMsg.java 6651900e79a5c3d4ad8329afbe3894544ce9f46e 
  ql/src/java/org/apache/hadoop/hive/ql/exec/DDLTask.java 87928ee930b5ee974d5e4144a584773a243f8d6f 
  ql/src/test/queries/clientnegative/parquet_alter_part_table_drop_columns.q PRE-CREATION 
  ql/src/test/results/clientnegative/parquet_alter_part_table_drop_columns.q.out PRE-CREATION 


Diff: https://reviews.apache.org/r/59446/diff/2/

Changes: https://reviews.apache.org/r/59446/diff/1-2/


Testing (updated)
-------

Added a negative qtest. Manually tested that no regression is caused for avro and textfile SerDes when columns are added or replaced in a partitioned table.


Thanks,

Barna Zsombor Klara