[jira] [Commented] (HIVE-11663) Auto load/unload custom udf function for hive cli and hiveserver2

2015-09-25 Thread Ashutosh Chauhan (JIRA)

[ 
https://issues.apache.org/jira/browse/HIVE-11663?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=14908481#comment-14908481
 ] 

Ashutosh Chauhan commented on HIVE-11663:
-

Permanent udfs introduced in HIVE-6047 is to serve the usecase which you have 
outlined. Wondering why using permanent udfs is not sufficient here?

> Auto load/unload custom udf function for hive cli and hiveserver2
> -
>
> Key: HIVE-11663
> URL: https://issues.apache.org/jira/browse/HIVE-11663
> Project: Hive
>  Issue Type: Improvement
>  Components: CLI, Configuration
>Affects Versions: 0.14.0, 1.0.0, 1.0.1, 1.1.1, 1.2.1
>Reporter: liuzongquan
>Assignee: liuzongquan
>  Labels: features, patch
> Attachments: HIVE-11663-2.patch
>
>   Original Estimate: 96h
>  Time Spent: 96h
>  Remaining Estimate: 0h
>
> when adding custom functions used in hiveserver2, the most method is re-build 
> the hive source code, re-dist and restart hiveserver2. This way will produce 
> big cost for service user and cluster manager. The best way, in my opinion, 
> the custom udf should be like a plugin to the hiveserver2 and hive cli, and  
> users can add and remove at run-time, especially for hiveserver2.



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)


[jira] [Commented] (HIVE-11663) Auto load/unload custom udf function for hive cli and hiveserver2

2015-09-17 Thread liuzongquan (JIRA)

[ 
https://issues.apache.org/jira/browse/HIVE-11663?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=14804894#comment-14804894
 ] 

liuzongquan commented on HIVE-11663:


Lefty,sorry for late~
I will change it as your suggestion.
Thank you so much!

> Auto load/unload custom udf function for hive cli and hiveserver2
> -
>
> Key: HIVE-11663
> URL: https://issues.apache.org/jira/browse/HIVE-11663
> Project: Hive
>  Issue Type: Improvement
>  Components: CLI, Configuration
>Affects Versions: 0.14.0, 1.0.0, 1.0.1, 1.1.1, 1.2.1
>Reporter: liuzongquan
>Assignee: liuzongquan
>  Labels: features, patch
> Attachments: HIVE-11663-2.patch
>
>   Original Estimate: 96h
>  Time Spent: 96h
>  Remaining Estimate: 0h
>
> when adding custom functions used in hiveserver2, the most method is re-build 
> the hive source code, re-dist and restart hiveserver2. This way will produce 
> big cost for service user and cluster manager. The best way, in my opinion, 
> the custom udf should be like a plugin to the hiveserver2 and hive cli, and  
> users can add and remove at run-time, especially for hiveserver2.



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)


[jira] [Commented] (HIVE-11663) Auto load/unload custom udf function for hive cli and hiveserver2

2015-09-05 Thread Lefty Leverenz (JIRA)

[ 
https://issues.apache.org/jira/browse/HIVE-11663?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=14731851#comment-14731851
 ] 

Lefty Leverenz commented on HIVE-11663:
---

[~liuzongquan@mls], sorry about the delay.  Here's my review of the 
configuration parameters in HiveConf.java, patch 2.

#  Why did you remove four comment lines starting with "// Hadoop Configuration 
Properties"?
#  Formatting:  For hive.server2.customized.udf.jars.path and 
hive.server2.customized.udf.properties please insert newlines (\n) in the 
descriptions as shown on lines 324 and 325, start the description of 
hive.server2.customized.udf.jars.path on a new line, and use standard indenting.
#  Capitalization:  For hive.server2.customized.udf.jars.path please change 
"The path on hdfs" to "The path on HDFS" and "means the hdfs path" to "means 
the HDFS path".
#  Edit:  For hive.server2.customized.udf.jars.path please change "When hdfs:// 
missing" to "When hdfs:// is missing".
#  Edits:  For hive.server2.customized.udf.properties please change "customized 
udf" to "customized UDF", change "defination" to "definition", change "It's 
must be under $" to "It must be under the $", and change "The patern is  as 
below." to "The pattern is".

A Review Board entry for this would make it easier to review, because comments 
could be posted on particular lines.


> Auto load/unload custom udf function for hive cli and hiveserver2
> -
>
> Key: HIVE-11663
> URL: https://issues.apache.org/jira/browse/HIVE-11663
> Project: Hive
>  Issue Type: Improvement
>  Components: CLI, Configuration
>Affects Versions: 0.14.0, 1.0.0, 1.0.1, 1.1.1, 1.2.1
>Reporter: liuzongquan
>Assignee: liuzongquan
>  Labels: features, patch
> Attachments: HIVE-11663-2.patch
>
>   Original Estimate: 96h
>  Time Spent: 96h
>  Remaining Estimate: 0h
>
> when adding custom functions used in hiveserver2, the most method is re-build 
> the hive source code, re-dist and restart hiveserver2. This way will produce 
> big cost for service user and cluster manager. The best way, in my opinion, 
> the custom udf should be like a plugin to the hiveserver2 and hive cli, and  
> users can add and remove at run-time, especially for hiveserver2.



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)


[jira] [Commented] (HIVE-11663) Auto load/unload custom udf function for hive cli and hiveserver2

2015-08-30 Thread Lefty Leverenz (JIRA)

[ 
https://issues.apache.org/jira/browse/HIVE-11663?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=14721766#comment-14721766
 ] 

Lefty Leverenz commented on HIVE-11663:
---

Thanks for the changes, I'll post more review comments soon.  But you need code 
review more than doc review, and I'm not qualified to review code.  Perhaps if 
you posted the patch on the review board you would get more responses.

* [Review Board | https://cwiki.apache.org/confluence/display/Hive/Review+Board]

If you have any trouble with the review board, I'll gladly help.  (I had 
trouble the first time, and finally discovered that I had to use a Firefox 
browser instead of Safari to get it to work.)

 Auto load/unload custom udf function for hive cli and hiveserver2
 -

 Key: HIVE-11663
 URL: https://issues.apache.org/jira/browse/HIVE-11663
 Project: Hive
  Issue Type: Improvement
  Components: CLI, Configuration
Affects Versions: 0.14.0, 1.0.0, 1.0.1, 1.1.1, 1.2.1
Reporter: liuzongquan
Assignee: liuzongquan
  Labels: features, patch
 Attachments: HIVE-11663-2.patch

   Original Estimate: 96h
  Time Spent: 96h
  Remaining Estimate: 0h

 when adding custom functions used in hiveserver2, the most method is re-build 
 the hive source code, re-dist and restart hiveserver2. This way will produce 
 big cost for service user and cluster manager. The best way, in my opinion, 
 the custom udf should be like a plugin to the hiveserver2 and hive cli, and  
 users can add and remove at run-time, especially for hiveserver2.



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)


[jira] [Commented] (HIVE-11663) Auto load/unload custom udf function for hive cli and hiveserver2

2015-08-27 Thread liuzongquan (JIRA)

[ 
https://issues.apache.org/jira/browse/HIVE-11663?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=14716148#comment-14716148
 ] 

liuzongquan commented on HIVE-11663:


I am a newer!
Thanks a lot, Lefty!
I'll modify the issue!

 Auto load/unload custom udf function for hive cli and hiveserver2
 -

 Key: HIVE-11663
 URL: https://issues.apache.org/jira/browse/HIVE-11663
 Project: Hive
  Issue Type: Improvement
  Components: CLI, Configuration
Affects Versions: 0.14.0, 1.0.0, 1.0.1, 1.1.1, 1.2.1
Reporter: liuzongquan
Assignee: liuzongquan
  Labels: features, patch
 Attachments: HIVE-11663.patch

   Original Estimate: 96h
  Remaining Estimate: 96h

 when adding custom functions used in hiveserver2, the most method is re-build 
 the hive source code, re-dist and restart hiveserver2. This way will produce 
 big cost for service user and cluster manager. The best way, in my opinion, 
 the custom udf should be like a plugin to the hiveserver2 and hive cli, and 
 users should be add and remove at run-time, especially for hiveserver2.



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)


[jira] [Commented] (HIVE-11663) Auto load/unload custom udf function for hive cli and hiveserver2

2015-08-27 Thread liuzongquan (JIRA)

[ 
https://issues.apache.org/jira/browse/HIVE-11663?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=14716150#comment-14716150
 ] 

liuzongquan commented on HIVE-11663:


I am a newer!
Thanks a lot, Lefty!
I'll reopen the issue!

 Auto load/unload custom udf function for hive cli and hiveserver2
 -

 Key: HIVE-11663
 URL: https://issues.apache.org/jira/browse/HIVE-11663
 Project: Hive
  Issue Type: Improvement
  Components: CLI, Configuration
Affects Versions: 0.14.0, 1.0.0, 1.0.1, 1.1.1, 1.2.1
Reporter: liuzongquan
Assignee: liuzongquan
  Labels: features, patch
 Attachments: HIVE-11663.patch

   Original Estimate: 96h
  Remaining Estimate: 96h

 when adding custom functions used in hiveserver2, the most method is re-build 
 the hive source code, re-dist and restart hiveserver2. This way will produce 
 big cost for service user and cluster manager. The best way, in my opinion, 
 the custom udf should be like a plugin to the hiveserver2 and hive cli, and 
 users should be add and remove at run-time, especially for hiveserver2.



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)


[jira] [Commented] (HIVE-11663) Auto load/unload custom udf function for hive cli and hiveserver2

2015-08-27 Thread liuzongquan (JIRA)

[ 
https://issues.apache.org/jira/browse/HIVE-11663?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=14716149#comment-14716149
 ] 

liuzongquan commented on HIVE-11663:


I am a newer!
Thanks a lot, Lefty!
I'll reopen the issue!

 Auto load/unload custom udf function for hive cli and hiveserver2
 -

 Key: HIVE-11663
 URL: https://issues.apache.org/jira/browse/HIVE-11663
 Project: Hive
  Issue Type: Improvement
  Components: CLI, Configuration
Affects Versions: 0.14.0, 1.0.0, 1.0.1, 1.1.1, 1.2.1
Reporter: liuzongquan
Assignee: liuzongquan
  Labels: features, patch
 Attachments: HIVE-11663.patch

   Original Estimate: 96h
  Remaining Estimate: 96h

 when adding custom functions used in hiveserver2, the most method is re-build 
 the hive source code, re-dist and restart hiveserver2. This way will produce 
 big cost for service user and cluster manager. The best way, in my opinion, 
 the custom udf should be like a plugin to the hiveserver2 and hive cli, and 
 users should be add and remove at run-time, especially for hiveserver2.



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)


[jira] [Commented] (HIVE-11663) Auto load/unload custom udf function for hive cli and hiveserver2

2015-08-27 Thread Lefty Leverenz (JIRA)

[ 
https://issues.apache.org/jira/browse/HIVE-11663?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=14716140#comment-14716140
 ] 

Lefty Leverenz commented on HIVE-11663:
---

This issue should not be marked Resolved until the patch has been reviewed, 
tested, and committed.  Please see How To Contribute:

* [How To Contribute -- Review Process | 
https://cwiki.apache.org/confluence/display/Hive/HowToContribute#HowToContribute-ReviewProcess]
* [How To Contribute -- Contributing Your Work | 
https://cwiki.apache.org/confluence/display/Hive/HowToContribute#HowToContribute-ContributingYourWork]

 Auto load/unload custom udf function for hive cli and hiveserver2
 -

 Key: HIVE-11663
 URL: https://issues.apache.org/jira/browse/HIVE-11663
 Project: Hive
  Issue Type: Improvement
  Components: CLI, Configuration
Affects Versions: 0.14.0, 1.0.0, 1.0.1, 1.1.1, 1.2.1
Reporter: liuzongquan
Assignee: liuzongquan
  Labels: features, patch
 Attachments: HIVE-11663.patch

   Original Estimate: 96h
  Remaining Estimate: 96h

 when adding custom functions used in hiveserver2, the most method is re-build 
 the hive source code, re-dist and restart hiveserver2. This way will produce 
 big cost for service user and cluster manager. The best way, in my opinion, 
 the custom udf should be like a plugin to the hiveserver2 and hive cli, and 
 users should be add and remove at run-time, especially for hiveserver2.



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)


[jira] [Commented] (HIVE-11663) Auto load/unload custom udf function for hive cli and hiveserver2

2015-08-27 Thread Lefty Leverenz (JIRA)

[ 
https://issues.apache.org/jira/browse/HIVE-11663?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=14716190#comment-14716190
 ] 

Lefty Leverenz commented on HIVE-11663:
---

Review comments for the first patch:

#  conf/hive-default.xml.template is now generated from HiveConf.java (see the 
WARNING!!! lines in the patch) so it doesn't belong in the patch
#  many or perhaps all files appear twice in the patch
#  new configuration parameters in HiveConf.java need editing:
#*   for *hive.server2.customized.udf.enabled* please change Whether enable 
hiveserver add customized udf from adding jars ... to Whether to enable 
HiveServer2 or the Hive CLI to add a customized UDF from a jar file ...
#*   for *hive.server2.customized.udf.jars.path* please supply a parameter 
description (even though it may seem obvious)
#*   for *hive.server2.customized.udf.properties* change The customized udf 
functions' defination file. It's must be under $HIVE_HOME/conf directory. The 
patern is  as below. to The customized UDF functions' definition file. It 
must be under the $HIVE_HOME/conf directory. The pattern is as below.
#*   for *hive.server2.customized.udf.defination.list* change defination to 
definition and for HIVE_SERVER2_CUSTOMIZED_UDF_DEFINATION_LIST similarly 
change DEFINATION to DEFINITION (also change all occurrences elsewhere in 
the patch -- I count 26 but that includes the template file and duplicate files)
#*   for the description of *hive.server2.customized.udf.defination.list* 
presumably the hive.thrift.customized.* parameters are supposed to be 
hive.server2.customized.* and other edits are needed: 
{code}
+   The customized UDF functions' definition list. For example, if 
there is a UDF jar on the \n +
+   ${hive.server2.customized.udf.jars.path} containing a foo.bar 
class which defines a UDF \n +
+   function named test, then you can add \'foo.bar:test\' to the 
property \n +
+   hive.server2.customized.udf.definition.list. Any additional 
definitions are separated with commas.\n +
+   Note that this property depends on 
hive.server2.customized.udf.enabled being set to true.
{code}



 Auto load/unload custom udf function for hive cli and hiveserver2
 -

 Key: HIVE-11663
 URL: https://issues.apache.org/jira/browse/HIVE-11663
 Project: Hive
  Issue Type: Improvement
  Components: CLI, Configuration
Affects Versions: 0.14.0, 1.0.0, 1.0.1, 1.1.1, 1.2.1
Reporter: liuzongquan
Assignee: liuzongquan
  Labels: features, patch
 Attachments: HIVE-11663.patch

   Original Estimate: 96h
  Remaining Estimate: 96h

 when adding custom functions used in hiveserver2, the most method is re-build 
 the hive source code, re-dist and restart hiveserver2. This way will produce 
 big cost for service user and cluster manager. The best way, in my opinion, 
 the custom udf should be like a plugin to the hiveserver2 and hive cli, and 
 users should be add and remove at run-time, especially for hiveserver2.



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)


[jira] [Commented] (HIVE-11663) Auto load/unload custom udf function for hive cli and hiveserver2

2015-08-27 Thread liuzongquan (JIRA)

[ 
https://issues.apache.org/jira/browse/HIVE-11663?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=14716231#comment-14716231
 ] 

liuzongquan commented on HIVE-11663:


I'll modify the patch as your suggest, Thank you so much, Lefty!

 Auto load/unload custom udf function for hive cli and hiveserver2
 -

 Key: HIVE-11663
 URL: https://issues.apache.org/jira/browse/HIVE-11663
 Project: Hive
  Issue Type: Improvement
  Components: CLI, Configuration
Affects Versions: 0.14.0, 1.0.0, 1.0.1, 1.1.1, 1.2.1
Reporter: liuzongquan
Assignee: liuzongquan
  Labels: features, patch
 Attachments: HIVE-11663.patch

   Original Estimate: 96h
  Remaining Estimate: 96h

 when adding custom functions used in hiveserver2, the most method is re-build 
 the hive source code, re-dist and restart hiveserver2. This way will produce 
 big cost for service user and cluster manager. The best way, in my opinion, 
 the custom udf should be like a plugin to the hiveserver2 and hive cli, and 
 users should be add and remove at run-time, especially for hiveserver2.



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)


[jira] [Commented] (HIVE-11663) Auto load/unload custom udf function for hive cli and hiveserver2

2015-08-26 Thread liuzongquan (JIRA)

[ 
https://issues.apache.org/jira/browse/HIVE-11663?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=14716072#comment-14716072
 ] 

liuzongquan commented on HIVE-11663:


The implementation is improve CLIService.java and CliDriver.java. When 
connecting hive throw thrift, CLIService will call function 
initCustomUDF(SessionHandle sessionHandle), this will load all functions 
configged in hive-udf.properties, which is a custom defined properties file 
underlying $HIVE_HOME/conf. Actully, hive-udf.properties can be renamed in 
hive-site.conf, which set all configuration for this feature, as below.

   property
namehive.server2.customized.udf.enabled/name
valuefalse/value
descriptionWhether enable hiveserver add customized udf from adding jars, 
default value is false/description
  /property
  property
namehive.server2.customized.udf.jars.path/name
value/
description/
  /property
  property
namehive.server2.customized.udf.properties/name
valuehive-udf.properties/value
descriptionThe customized udf functions' defination file. It's must be 
under $HIVE_HOME/conf directory. The patern is  as 
below.func_name={udf_jar_name,udf_class_name}/description
  /property




 Auto load/unload custom udf function for hive cli and hiveserver2
 -

 Key: HIVE-11663
 URL: https://issues.apache.org/jira/browse/HIVE-11663
 Project: Hive
  Issue Type: Improvement
  Components: CLI, Configuration
Affects Versions: 0.14.0, 1.0.0, 1.0.1, 1.1.1, 1.2.1
Reporter: liuzongquan
Assignee: liuzongquan
  Labels: features, patch
 Attachments: HIVE-11663.patch

   Original Estimate: 96h
  Remaining Estimate: 96h

 when adding custom functions used in hiveserver2, the most method is re-build 
 the hive source code, re-dist and restart hiveserver2. This way will produce 
 big cost for service user and cluster manager. The best way, in my opinion, 
 the custom udf should be like a plugin to the hiveserver2 and hive cli, and 
 users should be add and remove at run-time, especially for hiveserver2.



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)


[jira] [Commented] (HIVE-11663) Auto load/unload custom udf function for hive cli and hiveserver2

2015-08-26 Thread liuzongquan (JIRA)

[ 
https://issues.apache.org/jira/browse/HIVE-11663?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=14716080#comment-14716080
 ] 

liuzongquan commented on HIVE-11663:


{code:xml}
#  mls udf
multi_concat= {com.meilishuo.hive.udf.jar,com.meilishuo.hive.udf.UDFMultiConcat}
unfold={com.meilishuo.hive.udf.jar,com.meilishuo.hive.udf.UDFUnfoldURL}
ip2long={com.meilishuo.hive.udf.jar,com.meilishuo.hive.udf.UDFIp2Long}
{code}

 Auto load/unload custom udf function for hive cli and hiveserver2
 -

 Key: HIVE-11663
 URL: https://issues.apache.org/jira/browse/HIVE-11663
 Project: Hive
  Issue Type: Improvement
  Components: CLI, Configuration
Affects Versions: 0.14.0, 1.0.0, 1.0.1, 1.1.1, 1.2.1
Reporter: liuzongquan
Assignee: liuzongquan
  Labels: features, patch
 Attachments: HIVE-11663.patch

   Original Estimate: 96h
  Remaining Estimate: 96h

 when adding custom functions used in hiveserver2, the most method is re-build 
 the hive source code, re-dist and restart hiveserver2. This way will produce 
 big cost for service user and cluster manager. The best way, in my opinion, 
 the custom udf should be like a plugin to the hiveserver2 and hive cli, and 
 users should be add and remove at run-time, especially for hiveserver2.



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)


[jira] [Commented] (HIVE-11663) Auto load/unload custom udf function for hive cli and hiveserver2

2015-08-26 Thread liuzongquan (JIRA)

[ 
https://issues.apache.org/jira/browse/HIVE-11663?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=14716081#comment-14716081
 ] 

liuzongquan commented on HIVE-11663:


{code:xml}
## hive-site.xml 
property
namehive.server2.customized.udf.enabled/name
valuefalse/value
descriptionWhether enable hiveserver add customized udf from adding jars, 
default value is false/description
/property
property
namehive.server2.customized.udf.jars.path/name
value/
description/
/property
property
namehive.server2.customized.udf.properties/name
valuehive-udf.properties/value
descriptionThe customized udf functions' defination file. It's must be under 
$HIVE_HOME/conf directory. The patern is as 
below.func_name={udf_jar_name,udf_class_name}
/description
/property
{code}

 Auto load/unload custom udf function for hive cli and hiveserver2
 -

 Key: HIVE-11663
 URL: https://issues.apache.org/jira/browse/HIVE-11663
 Project: Hive
  Issue Type: Improvement
  Components: CLI, Configuration
Affects Versions: 0.14.0, 1.0.0, 1.0.1, 1.1.1, 1.2.1
Reporter: liuzongquan
Assignee: liuzongquan
  Labels: features, patch
 Attachments: HIVE-11663.patch

   Original Estimate: 96h
  Remaining Estimate: 96h

 when adding custom functions used in hiveserver2, the most method is re-build 
 the hive source code, re-dist and restart hiveserver2. This way will produce 
 big cost for service user and cluster manager. The best way, in my opinion, 
 the custom udf should be like a plugin to the hiveserver2 and hive cli, and 
 users should be add and remove at run-time, especially for hiveserver2.



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)


[jira] [Commented] (HIVE-11663) Auto load/unload custom udf function for hive cli and hiveserver2

2015-08-26 Thread liuzongquan (JIRA)

[ 
https://issues.apache.org/jira/browse/HIVE-11663?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=14716077#comment-14716077
 ] 

liuzongquan commented on HIVE-11663:


the right style for hive-udf.properties:

# mls udf
multi_concat= {com.meilishuo.hive.udf.jar,com.meilishuo.hive.udf.UDFMultiConcat}
unfold={com.meilishuo.hive.udf.jar,com.meilishuo.hive.udf.UDFUnfoldURL}
ip2long={com.meilishuo.hive.udf.jar,com.meilishuo.hive.udf.UDFIp2Long}

 Auto load/unload custom udf function for hive cli and hiveserver2
 -

 Key: HIVE-11663
 URL: https://issues.apache.org/jira/browse/HIVE-11663
 Project: Hive
  Issue Type: Improvement
  Components: CLI, Configuration
Affects Versions: 0.14.0, 1.0.0, 1.0.1, 1.1.1, 1.2.1
Reporter: liuzongquan
Assignee: liuzongquan
  Labels: features, patch
 Attachments: HIVE-11663.patch

   Original Estimate: 96h
  Remaining Estimate: 96h

 when adding custom functions used in hiveserver2, the most method is re-build 
 the hive source code, re-dist and restart hiveserver2. This way will produce 
 big cost for service user and cluster manager. The best way, in my opinion, 
 the custom udf should be like a plugin to the hiveserver2 and hive cli, and 
 users should be add and remove at run-time, especially for hiveserver2.



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)


[jira] [Commented] (HIVE-11663) Auto load/unload custom udf function for hive cli and hiveserver2

2015-08-26 Thread liuzongquan (JIRA)

[ 
https://issues.apache.org/jira/browse/HIVE-11663?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=14716076#comment-14716076
 ] 

liuzongquan commented on HIVE-11663:


when add a custom udf that want to be used in hiveserver2 and hive cli, you 
just upload the UDF jar file on to ${hive.server2.customized.udf.jar.path}, 
which is a hdfs:// path. And then, append the udf properties to 
hive-udf.properties as following.
 # mls udf
 multi_concat={com.meilishuo.hive.udf.jar,com.meilishuo.hive.udf.UDFMultiConcat}
 unfold={com.meilishuo.hive.udf.jar,com.meilishuo.hive.udf.UDFUnfoldURL}
 ip2long={com.meilishuo.hive.udf.jar,com.meilishuo.hive.udf.UDFIp2Long}

That's all you do for new custom UDF! And the hiveserver2 can use them 
immediately!

 Auto load/unload custom udf function for hive cli and hiveserver2
 -

 Key: HIVE-11663
 URL: https://issues.apache.org/jira/browse/HIVE-11663
 Project: Hive
  Issue Type: Improvement
  Components: CLI, Configuration
Affects Versions: 0.14.0, 1.0.0, 1.0.1, 1.1.1, 1.2.1
Reporter: liuzongquan
Assignee: liuzongquan
  Labels: features, patch
 Attachments: HIVE-11663.patch

   Original Estimate: 96h
  Remaining Estimate: 96h

 when adding custom functions used in hiveserver2, the most method is re-build 
 the hive source code, re-dist and restart hiveserver2. This way will produce 
 big cost for service user and cluster manager. The best way, in my opinion, 
 the custom udf should be like a plugin to the hiveserver2 and hive cli, and 
 users should be add and remove at run-time, especially for hiveserver2.



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)