You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@hive.apache.org by "Namit Jain (JIRA)" <ji...@apache.org> on 2009/07/29 20:54:15 UTC

[jira] Created: (HIVE-708) for tranform,let the user specify the serde

for tranform,let the user specify the serde
-------------------------------------------

                 Key: HIVE-708
                 URL: https://issues.apache.org/jira/browse/HIVE-708
             Project: Hadoop Hive
          Issue Type: New Feature
          Components: Serializers/Deserializers
            Reporter: Namit Jain
            Assignee: Namit Jain


Currently, LazySimpleSerDe is used to send data to the user transformation functions - it would be useful to let the user specify the format of the data.

Specifically, it would be very easy and useful to accommodate: 


(cut and paste from Venky's mail)


Here's the typedbytes stuff that Dumbo uses.

http://issues.apache.org/jira/browse/HADOOP-1722

From:
http://static.last.fm/johan/huguk-20090414/klaas-hadoop-1722.pdf
Timings for IP count program on 300gigs of weblogs:
Java: 8 minutes
Dumbo with typed bytes: 10 minutes
Hive: 13 minutes
Dumbo without typed bytes: 16 minutes

They also have a fast python decoder for this, which is apparently 25% faster than the python version. 
http://github.com/klbostee/ctypedbytes/tree/master

http://dumbotics.com/2009/05/31/dumbo-on-clouderas-distribution/

-- 
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.


[jira] Assigned: (HIVE-708) for tranform,let the user specify the serde

Posted by "Zheng Shao (JIRA)" <ji...@apache.org>.
     [ https://issues.apache.org/jira/browse/HIVE-708?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Zheng Shao reassigned HIVE-708:
-------------------------------

    Assignee: Zheng Shao  (was: Namit Jain)

> for tranform,let the user specify the serde
> -------------------------------------------
>
>                 Key: HIVE-708
>                 URL: https://issues.apache.org/jira/browse/HIVE-708
>             Project: Hadoop Hive
>          Issue Type: New Feature
>          Components: Serializers/Deserializers
>            Reporter: Namit Jain
>            Assignee: Zheng Shao
>         Attachments: hive.708.1.patch
>
>
> Currently, LazySimpleSerDe is used to send data to the user transformation functions - it would be useful to let the user specify the format of the data.
> Specifically, it would be very easy and useful to accommodate: 
> (cut and paste from Venky's mail)
> Here's the typedbytes stuff that Dumbo uses.
> http://issues.apache.org/jira/browse/HADOOP-1722
> From:
> http://static.last.fm/johan/huguk-20090414/klaas-hadoop-1722.pdf
> Timings for IP count program on 300gigs of weblogs:
> Java: 8 minutes
> Dumbo with typed bytes: 10 minutes
> Hive: 13 minutes
> Dumbo without typed bytes: 16 minutes
> They also have a fast python decoder for this, which is apparently 25% faster than the python version. 
> http://github.com/klbostee/ctypedbytes/tree/master
> http://dumbotics.com/2009/05/31/dumbo-on-clouderas-distribution/

-- 
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.


[jira] Updated: (HIVE-708) for tranform,let the user specify the serde

Posted by "Namit Jain (JIRA)" <ji...@apache.org>.
     [ https://issues.apache.org/jira/browse/HIVE-708?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Namit Jain updated HIVE-708:
----------------------------

    Attachment: hive.708.1.patch

> for tranform,let the user specify the serde
> -------------------------------------------
>
>                 Key: HIVE-708
>                 URL: https://issues.apache.org/jira/browse/HIVE-708
>             Project: Hadoop Hive
>          Issue Type: New Feature
>          Components: Serializers/Deserializers
>            Reporter: Namit Jain
>            Assignee: Namit Jain
>         Attachments: hive.708.1.patch
>
>
> Currently, LazySimpleSerDe is used to send data to the user transformation functions - it would be useful to let the user specify the format of the data.
> Specifically, it would be very easy and useful to accommodate: 
> (cut and paste from Venky's mail)
> Here's the typedbytes stuff that Dumbo uses.
> http://issues.apache.org/jira/browse/HADOOP-1722
> From:
> http://static.last.fm/johan/huguk-20090414/klaas-hadoop-1722.pdf
> Timings for IP count program on 300gigs of weblogs:
> Java: 8 minutes
> Dumbo with typed bytes: 10 minutes
> Hive: 13 minutes
> Dumbo without typed bytes: 16 minutes
> They also have a fast python decoder for this, which is apparently 25% faster than the python version. 
> http://github.com/klbostee/ctypedbytes/tree/master
> http://dumbotics.com/2009/05/31/dumbo-on-clouderas-distribution/

-- 
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.


[jira] Commented: (HIVE-708) for tranform,let the user specify the serde

Posted by "Zheng Shao (JIRA)" <ji...@apache.org>.
    [ https://issues.apache.org/jira/browse/HIVE-708?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12736904#action_12736904 ] 

Zheng Shao commented on HIVE-708:
---------------------------------

Can we also extract "row format..." to be a common sub tree and have a utility function processing it (which returns a SerDe), so that create table and transform can reuse (in the future, load, and plain select can support this as well).

> for tranform,let the user specify the serde
> -------------------------------------------
>
>                 Key: HIVE-708
>                 URL: https://issues.apache.org/jira/browse/HIVE-708
>             Project: Hadoop Hive
>          Issue Type: New Feature
>          Components: Serializers/Deserializers
>            Reporter: Namit Jain
>            Assignee: Namit Jain
>         Attachments: hive.708.1.patch
>
>
> Currently, LazySimpleSerDe is used to send data to the user transformation functions - it would be useful to let the user specify the format of the data.
> Specifically, it would be very easy and useful to accommodate: 
> (cut and paste from Venky's mail)
> Here's the typedbytes stuff that Dumbo uses.
> http://issues.apache.org/jira/browse/HADOOP-1722
> From:
> http://static.last.fm/johan/huguk-20090414/klaas-hadoop-1722.pdf
> Timings for IP count program on 300gigs of weblogs:
> Java: 8 minutes
> Dumbo with typed bytes: 10 minutes
> Hive: 13 minutes
> Dumbo without typed bytes: 16 minutes
> They also have a fast python decoder for this, which is apparently 25% faster than the python version. 
> http://github.com/klbostee/ctypedbytes/tree/master
> http://dumbotics.com/2009/05/31/dumbo-on-clouderas-distribution/

-- 
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.


[jira] Commented: (HIVE-708) Add TypedBytes SerDe for transform

Posted by "Zheng Shao (JIRA)" <ji...@apache.org>.
    [ https://issues.apache.org/jira/browse/HIVE-708?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12739151#action_12739151 ] 

Zheng Shao commented on HIVE-708:
---------------------------------

Looked at the code a bit more. It seems it's compiled against 0.19.  We need to make the following changes to make it compile under 0.17:
* Replace org.apache.hadoop.io.ByteWritable with org.apache.hadoop.hive.serde2.io.ByteWritable. (The former is not in hadoop 0.17)
* Change the calls to org.apache.hadoop.io.BytesWritable - only get() and getSize() are available in hadoop 0.17.


> Add TypedBytes SerDe for transform
> ----------------------------------
>
>                 Key: HIVE-708
>                 URL: https://issues.apache.org/jira/browse/HIVE-708
>             Project: Hadoop Hive
>          Issue Type: New Feature
>          Components: Serializers/Deserializers
>            Reporter: Namit Jain
>            Assignee: Namit Jain
>         Attachments: hive.708.1.patch
>
>
> Currently, LazySimpleSerDe is used to send data to the user transformation functions - it would be useful to let the user specify the format of the data.
> Specifically, it would be very easy and useful to accommodate: 
> (cut and paste from Venky's mail)
> Here's the typedbytes stuff that Dumbo uses.
> http://issues.apache.org/jira/browse/HADOOP-1722
> From:
> http://static.last.fm/johan/huguk-20090414/klaas-hadoop-1722.pdf
> Timings for IP count program on 300gigs of weblogs:
> Java: 8 minutes
> Dumbo with typed bytes: 10 minutes
> Hive: 13 minutes
> Dumbo without typed bytes: 16 minutes
> They also have a fast python decoder for this, which is apparently 25% faster than the python version. 
> http://github.com/klbostee/ctypedbytes/tree/master
> http://dumbotics.com/2009/05/31/dumbo-on-clouderas-distribution/

-- 
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.


[jira] Commented: (HIVE-708) for tranform,let the user specify the serde

Posted by "Namit Jain (JIRA)" <ji...@apache.org>.
    [ https://issues.apache.org/jira/browse/HIVE-708?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12736840#action_12736840 ] 

Namit Jain commented on HIVE-708:
---------------------------------

Sure- I will make that change

> for tranform,let the user specify the serde
> -------------------------------------------
>
>                 Key: HIVE-708
>                 URL: https://issues.apache.org/jira/browse/HIVE-708
>             Project: Hadoop Hive
>          Issue Type: New Feature
>          Components: Serializers/Deserializers
>            Reporter: Namit Jain
>            Assignee: Namit Jain
>         Attachments: hive.708.1.patch
>
>
> Currently, LazySimpleSerDe is used to send data to the user transformation functions - it would be useful to let the user specify the format of the data.
> Specifically, it would be very easy and useful to accommodate: 
> (cut and paste from Venky's mail)
> Here's the typedbytes stuff that Dumbo uses.
> http://issues.apache.org/jira/browse/HADOOP-1722
> From:
> http://static.last.fm/johan/huguk-20090414/klaas-hadoop-1722.pdf
> Timings for IP count program on 300gigs of weblogs:
> Java: 8 minutes
> Dumbo with typed bytes: 10 minutes
> Hive: 13 minutes
> Dumbo without typed bytes: 16 minutes
> They also have a fast python decoder for this, which is apparently 25% faster than the python version. 
> http://github.com/klbostee/ctypedbytes/tree/master
> http://dumbotics.com/2009/05/31/dumbo-on-clouderas-distribution/

-- 
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.


[jira] Commented: (HIVE-708) for tranform,let the user specify the serde

Posted by "Zheng Shao (JIRA)" <ji...@apache.org>.
    [ https://issues.apache.org/jira/browse/HIVE-708?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12736835#action_12736835 ] 

Zheng Shao commented on HIVE-708:
---------------------------------

@hive.708.1.patch:

For the grammar change, can we accept SerDe at both input and output of the user script?

{code}
SELECT TRANSFORM(a, b, c)
ROW FORMAT DELIMITED FIELDS TERMINATED BY '\001'
USING 'mypython.py'
AS (d STRING, e INT, f DOUBLE)
ROW FORMAT DELIMITED FIELDS TERMINATED BY '\t'
{code}


> for tranform,let the user specify the serde
> -------------------------------------------
>
>                 Key: HIVE-708
>                 URL: https://issues.apache.org/jira/browse/HIVE-708
>             Project: Hadoop Hive
>          Issue Type: New Feature
>          Components: Serializers/Deserializers
>            Reporter: Namit Jain
>            Assignee: Namit Jain
>         Attachments: hive.708.1.patch
>
>
> Currently, LazySimpleSerDe is used to send data to the user transformation functions - it would be useful to let the user specify the format of the data.
> Specifically, it would be very easy and useful to accommodate: 
> (cut and paste from Venky's mail)
> Here's the typedbytes stuff that Dumbo uses.
> http://issues.apache.org/jira/browse/HADOOP-1722
> From:
> http://static.last.fm/johan/huguk-20090414/klaas-hadoop-1722.pdf
> Timings for IP count program on 300gigs of weblogs:
> Java: 8 minutes
> Dumbo with typed bytes: 10 minutes
> Hive: 13 minutes
> Dumbo without typed bytes: 16 minutes
> They also have a fast python decoder for this, which is apparently 25% faster than the python version. 
> http://github.com/klbostee/ctypedbytes/tree/master
> http://dumbotics.com/2009/05/31/dumbo-on-clouderas-distribution/

-- 
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.


[jira] Updated: (HIVE-708) Add TypedBytes SerDe for transform

Posted by "Namit Jain (JIRA)" <ji...@apache.org>.
     [ https://issues.apache.org/jira/browse/HIVE-708?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Namit Jain updated HIVE-708:
----------------------------

    Attachment: hive.708.2.patch

> Add TypedBytes SerDe for transform
> ----------------------------------
>
>                 Key: HIVE-708
>                 URL: https://issues.apache.org/jira/browse/HIVE-708
>             Project: Hadoop Hive
>          Issue Type: New Feature
>          Components: Serializers/Deserializers
>            Reporter: Namit Jain
>            Assignee: Namit Jain
>         Attachments: hive.708.1.patch, hive.708.2.patch
>
>
> Currently, LazySimpleSerDe is used to send data to the user transformation functions - it would be useful to let the user specify the format of the data.
> Specifically, it would be very easy and useful to accommodate: 
> (cut and paste from Venky's mail)
> Here's the typedbytes stuff that Dumbo uses.
> http://issues.apache.org/jira/browse/HADOOP-1722
> From:
> http://static.last.fm/johan/huguk-20090414/klaas-hadoop-1722.pdf
> Timings for IP count program on 300gigs of weblogs:
> Java: 8 minutes
> Dumbo with typed bytes: 10 minutes
> Hive: 13 minutes
> Dumbo without typed bytes: 16 minutes
> They also have a fast python decoder for this, which is apparently 25% faster than the python version. 
> http://github.com/klbostee/ctypedbytes/tree/master
> http://dumbotics.com/2009/05/31/dumbo-on-clouderas-distribution/

-- 
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.


[jira] Updated: (HIVE-708) Add TypedBytes SerDe for transform

Posted by "Namit Jain (JIRA)" <ji...@apache.org>.
     [ https://issues.apache.org/jira/browse/HIVE-708?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Namit Jain updated HIVE-708:
----------------------------

    Attachment: hive.708.3.patch

> Add TypedBytes SerDe for transform
> ----------------------------------
>
>                 Key: HIVE-708
>                 URL: https://issues.apache.org/jira/browse/HIVE-708
>             Project: Hadoop Hive
>          Issue Type: New Feature
>          Components: Serializers/Deserializers
>            Reporter: Namit Jain
>            Assignee: Namit Jain
>         Attachments: hive.708.1.patch, hive.708.2.patch, hive.708.3.patch
>
>
> Currently, LazySimpleSerDe is used to send data to the user transformation functions - it would be useful to let the user specify the format of the data.
> Specifically, it would be very easy and useful to accommodate: 
> (cut and paste from Venky's mail)
> Here's the typedbytes stuff that Dumbo uses.
> http://issues.apache.org/jira/browse/HADOOP-1722
> From:
> http://static.last.fm/johan/huguk-20090414/klaas-hadoop-1722.pdf
> Timings for IP count program on 300gigs of weblogs:
> Java: 8 minutes
> Dumbo with typed bytes: 10 minutes
> Hive: 13 minutes
> Dumbo without typed bytes: 16 minutes
> They also have a fast python decoder for this, which is apparently 25% faster than the python version. 
> http://github.com/klbostee/ctypedbytes/tree/master
> http://dumbotics.com/2009/05/31/dumbo-on-clouderas-distribution/

-- 
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.


[jira] Commented: (HIVE-708) Add TypedBytes SerDe for transform

Posted by "Klaas Bosteels (JIRA)" <ji...@apache.org>.
    [ https://issues.apache.org/jira/browse/HIVE-708?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12739387#action_12739387 ] 

Klaas Bosteels commented on HIVE-708:
-------------------------------------

FWIW, the patch seems to be affected by MAPREDUCE-764.

> Add TypedBytes SerDe for transform
> ----------------------------------
>
>                 Key: HIVE-708
>                 URL: https://issues.apache.org/jira/browse/HIVE-708
>             Project: Hadoop Hive
>          Issue Type: New Feature
>          Components: Serializers/Deserializers
>            Reporter: Namit Jain
>            Assignee: Namit Jain
>         Attachments: hive.708.1.patch
>
>
> Currently, LazySimpleSerDe is used to send data to the user transformation functions - it would be useful to let the user specify the format of the data.
> Specifically, it would be very easy and useful to accommodate: 
> (cut and paste from Venky's mail)
> Here's the typedbytes stuff that Dumbo uses.
> http://issues.apache.org/jira/browse/HADOOP-1722
> From:
> http://static.last.fm/johan/huguk-20090414/klaas-hadoop-1722.pdf
> Timings for IP count program on 300gigs of weblogs:
> Java: 8 minutes
> Dumbo with typed bytes: 10 minutes
> Hive: 13 minutes
> Dumbo without typed bytes: 16 minutes
> They also have a fast python decoder for this, which is apparently 25% faster than the python version. 
> http://github.com/klbostee/ctypedbytes/tree/master
> http://dumbotics.com/2009/05/31/dumbo-on-clouderas-distribution/

-- 
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.


[jira] Updated: (HIVE-708) Add TypedBytes SerDe for transform

Posted by "Zheng Shao (JIRA)" <ji...@apache.org>.
     [ https://issues.apache.org/jira/browse/HIVE-708?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Zheng Shao updated HIVE-708:
----------------------------

       Resolution: Fixed
    Fix Version/s: 0.5.0
     Release Note: HIVE-708. Add TypedBytes SerDe for transform. (Namit Jain via zshao)
     Hadoop Flags: [Reviewed]
           Status: Resolved  (was: Patch Available)

Committed. Thanks Namit.
Please take a look at the follow-ups when you have time.


> Add TypedBytes SerDe for transform
> ----------------------------------
>
>                 Key: HIVE-708
>                 URL: https://issues.apache.org/jira/browse/HIVE-708
>             Project: Hadoop Hive
>          Issue Type: New Feature
>          Components: Serializers/Deserializers
>            Reporter: Namit Jain
>            Assignee: Namit Jain
>             Fix For: 0.5.0
>
>         Attachments: hive.708.1.patch, hive.708.2.patch, hive.708.3.patch, hive.708.4.patch
>
>
> Currently, LazySimpleSerDe is used to send data to the user transformation functions - it would be useful to let the user specify the format of the data.
> Specifically, it would be very easy and useful to accommodate: 
> (cut and paste from Venky's mail)
> Here's the typedbytes stuff that Dumbo uses.
> http://issues.apache.org/jira/browse/HADOOP-1722
> From:
> http://static.last.fm/johan/huguk-20090414/klaas-hadoop-1722.pdf
> Timings for IP count program on 300gigs of weblogs:
> Java: 8 minutes
> Dumbo with typed bytes: 10 minutes
> Hive: 13 minutes
> Dumbo without typed bytes: 16 minutes
> They also have a fast python decoder for this, which is apparently 25% faster than the python version. 
> http://github.com/klbostee/ctypedbytes/tree/master
> http://dumbotics.com/2009/05/31/dumbo-on-clouderas-distribution/

-- 
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.


[jira] Updated: (HIVE-708) Add TypedBytes SerDe for transform

Posted by "Zheng Shao (JIRA)" <ji...@apache.org>.
     [ https://issues.apache.org/jira/browse/HIVE-708?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Zheng Shao updated HIVE-708:
----------------------------

    Assignee: Namit Jain  (was: Zheng Shao)
     Summary: Add TypedBytes SerDe for transform  (was: for tranform,let the user specify the serde)

> Add TypedBytes SerDe for transform
> ----------------------------------
>
>                 Key: HIVE-708
>                 URL: https://issues.apache.org/jira/browse/HIVE-708
>             Project: Hadoop Hive
>          Issue Type: New Feature
>          Components: Serializers/Deserializers
>            Reporter: Namit Jain
>            Assignee: Namit Jain
>         Attachments: hive.708.1.patch
>
>
> Currently, LazySimpleSerDe is used to send data to the user transformation functions - it would be useful to let the user specify the format of the data.
> Specifically, it would be very easy and useful to accommodate: 
> (cut and paste from Venky's mail)
> Here's the typedbytes stuff that Dumbo uses.
> http://issues.apache.org/jira/browse/HADOOP-1722
> From:
> http://static.last.fm/johan/huguk-20090414/klaas-hadoop-1722.pdf
> Timings for IP count program on 300gigs of weblogs:
> Java: 8 minutes
> Dumbo with typed bytes: 10 minutes
> Hive: 13 minutes
> Dumbo without typed bytes: 16 minutes
> They also have a fast python decoder for this, which is apparently 25% faster than the python version. 
> http://github.com/klbostee/ctypedbytes/tree/master
> http://dumbotics.com/2009/05/31/dumbo-on-clouderas-distribution/

-- 
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.


[jira] Commented: (HIVE-708) Add TypedBytes SerDe for transform

Posted by "Zheng Shao (JIRA)" <ji...@apache.org>.
    [ https://issues.apache.org/jira/browse/HIVE-708?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12747755#action_12747755 ] 

Zheng Shao commented on HIVE-708:
---------------------------------

The new test failed somehow. I had to revert the new test for now.


> Add TypedBytes SerDe for transform
> ----------------------------------
>
>                 Key: HIVE-708
>                 URL: https://issues.apache.org/jira/browse/HIVE-708
>             Project: Hadoop Hive
>          Issue Type: New Feature
>          Components: Serializers/Deserializers
>            Reporter: Namit Jain
>            Assignee: Namit Jain
>             Fix For: 0.5.0
>
>         Attachments: hive.708.1.patch, hive.708.2.patch, hive.708.3.patch, hive.708.4.patch
>
>
> Currently, LazySimpleSerDe is used to send data to the user transformation functions - it would be useful to let the user specify the format of the data.
> Specifically, it would be very easy and useful to accommodate: 
> (cut and paste from Venky's mail)
> Here's the typedbytes stuff that Dumbo uses.
> http://issues.apache.org/jira/browse/HADOOP-1722
> From:
> http://static.last.fm/johan/huguk-20090414/klaas-hadoop-1722.pdf
> Timings for IP count program on 300gigs of weblogs:
> Java: 8 minutes
> Dumbo with typed bytes: 10 minutes
> Hive: 13 minutes
> Dumbo without typed bytes: 16 minutes
> They also have a fast python decoder for this, which is apparently 25% faster than the python version. 
> http://github.com/klbostee/ctypedbytes/tree/master
> http://dumbotics.com/2009/05/31/dumbo-on-clouderas-distribution/

-- 
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.