RE: Impact of partitioning on certain queries

[Mich Talebzadeh]

Well that is debatable.

 

The following table sales is partitioned in Oracle but has local bitmap indexes 
that help the query.

 

select * from sales where prod_id = 10;

 

no rows selected

 

 

Execution Plan

--

Plan hash value: 511273406

 

-

| Id  | Operation  | Name   | Rows  | Bytes | 
Cost (%CPU)| Time | Pstart| Pstop |

-

|   0 | SELECT STATEMENT   ||   347 | 10063 |   
 93   (0)| 00:00:02 |   |   |

|   1 |  PARTITION RANGE ALL   ||   347 | 10063 |   
 93   (0)| 00:00:02 | 1 |28 |

|   2 |   TABLE ACCESS BY LOCAL INDEX ROWID| SALES  |   347 | 10063 |   
 93   (0)| 00:00:02 | 1 |28 |

|   3 |BITMAP CONVERSION TO ROWIDS ||   |   |   
 |  |   |   |

|*  4 | BITMAP INDEX SINGLE VALUE  | SALES_PROD_BIX |   |   |   
 |  | 1 |28 |

-

 

Obviously at this stage we do not have local indexes in Hive. That could make 
it moredefficient for search and IMO will be a great tool.

 

Cheers,

 

 

Dr Mich Talebzadeh

 

LinkedIn   
<https://www.linkedin.com/profile/view?id=AAEWh2gBxianrbJd6zP6AcPCCdOABUrV8Pw>
 
https://www.linkedin.com/profile/view?id=AAEWh2gBxianrbJd6zP6AcPCCdOABUrV8Pw

 

Sybase ASE 15 Gold Medal Award 2008

A Winning Strategy: Running the most Critical Financial Data on ASE 15

 
<http://login.sybase.com/files/Product_Overviews/ASE-Winning-Strategy-091908.pdf>
 http://login.sybase.com/files/Product_Overviews/ASE-Winning-Strategy-091908.pdf

Author of the books "A Practitioner’s Guide to Upgrading to Sybase ASE 15", 
ISBN 978-0-9563693-0-7. 

co-author "Sybase Transact SQL Guidelines Best Practices", ISBN 
978-0-9759693-0-4

Publications due shortly:

Complex Event Processing in Heterogeneous Environments, ISBN: 978-0-9563693-3-8

Oracle and Sybase, Concepts and Contrasts, ISBN: 978-0-9563693-1-4, volume one 
out shortly

 

 <http://talebzadehmich.wordpress.com/> http://talebzadehmich.wordpress.com

 

NOTE: The information in this email is proprietary and confidential. This 
message is for the designated recipient only, if you are not the intended 
recipient, you should destroy it immediately. Any information in this message 
shall not be understood as given or endorsed by Peridale Technology Ltd, its 
subsidiaries or their employees, unless expressly so stated. It is the 
responsibility of the recipient to ensure that this email is virus free, 
therefore neither Peridale Ltd, its subsidiaries nor their employees accept any 
responsibility.

 

From: Jörn Franke [mailto:jornfra...@gmail.com] 
Sent: 08 January 2016 06:20
To: user@hive.apache.org
Subject: Re: Impact of partitioning on certain queries

 

This observation is correct and it is the same  behavior as you see it in other 
databases supporting partitions. Usually you should avoid many small partitions.


On 07 Jan 2016, at 23:53, Mich Talebzadeh <m...@peridale.co.uk 
<mailto:m...@peridale.co.uk> > wrote:

Ok we hope that partitioning improves performance where the predicate is on 
partitioned columns

 

I have two tables. One a basic table called smallsales defined as below

 

CREATE TABLE `smallsales`(  |

|   `prod_id` bigint, |

|   `cust_id` bigint, |

|   `time_id` timestamp,  |

|   `channel_id` bigint,  |

|   `promo_id` bigint,|

|   `quantity_sold` decimal(10,0),|

|   `amount_sold` decimal(10,0))  |

| ROW FORMAT SERDE|

|   'org.apache.hadoop.hive.serde2.lazy.LazySimpleSerDe'  |

| STORED AS INPUTFORMAT   |

|   'org.apache.hadoop.mapred.TextInputFormat'|

| OUTPUTFORMAT|

|   'org.apache.hadoop.hive.ql.io.HiveIgnoreKeyTextOutputFormat'  |

| LOCATION|

|   'hdfs://rhes564:9000/user/hive/warehouse/oraclehadoop.db/smallsales'  |

| TBLPROPERTIES (   

Re: Impact of partitioning on certain queries

[Jörn Franke]

Well you use a text format for your data so you should not be surprised. For 
text based formats, such as csv, you can always use the hive bitmap index. I do 
not think it makes a lot of sense to compare here processing csv files and 
internal tables of a relational database.

> On 08 Jan 2016, at 09:30, Mich Talebzadeh <m...@peridale.co.uk> wrote:
> 
> Well that is debatable.
>  
> The following table sales is partitioned in Oracle but has local bitmap 
> indexes that help the query.
>  
> select * from sales where prod_id = 10;
>  
> no rows selected
>  
>  
> Execution Plan
> --
> Plan hash value: 511273406
>  
> -
> | Id  | Operation  | Name   | Rows  | Bytes | 
> Cost (%CPU)| Time | Pstart| Pstop |
> -
> |   0 | SELECT STATEMENT   ||   347 | 10063 | 
>93   (0)| 00:00:02 |   |   |
> |   1 |  PARTITION RANGE ALL   ||   347 | 10063 | 
>93   (0)| 00:00:02 | 1 |28 |
> |   2 |   TABLE ACCESS BY LOCAL INDEX ROWID| SALES  |   347 | 10063 | 
>93   (0)| 00:00:02 | 1 |28 |
> |   3 |BITMAP CONVERSION TO ROWIDS ||   |   | 
>|  |   |   |
> |*  4 | BITMAP INDEX SINGLE VALUE  | SALES_PROD_BIX |   |   | 
>|  | 1 |28 |
> -
>  
> Obviously at this stage we do not have local indexes in Hive. That could make 
> it moredefficient for search and IMO will be a great tool.
>  
> Cheers,
>  
>  
> Dr Mich Talebzadeh
>  
> LinkedIn  
> https://www.linkedin.com/profile/view?id=AAEWh2gBxianrbJd6zP6AcPCCdOABUrV8Pw
>  
> Sybase ASE 15 Gold Medal Award 2008
> A Winning Strategy: Running the most Critical Financial Data on ASE 15
> http://login.sybase.com/files/Product_Overviews/ASE-Winning-Strategy-091908.pdf
> Author of the books "A Practitioner’s Guide to Upgrading to Sybase ASE 15", 
> ISBN 978-0-9563693-0-7.
> co-author "Sybase Transact SQL Guidelines Best Practices", ISBN 
> 978-0-9759693-0-4
> Publications due shortly:
> Complex Event Processing in Heterogeneous Environments, ISBN: 
> 978-0-9563693-3-8
> Oracle and Sybase, Concepts and Contrasts, ISBN: 978-0-9563693-1-4, volume 
> one out shortly
>  
> http://talebzadehmich.wordpress.com
>  
> NOTE: The information in this email is proprietary and confidential. This 
> message is for the designated recipient only, if you are not the intended 
> recipient, you should destroy it immediately. Any information in this message 
> shall not be understood as given or endorsed by Peridale Technology Ltd, its 
> subsidiaries or their employees, unless expressly so stated. It is the 
> responsibility of the recipient to ensure that this email is virus free, 
> therefore neither Peridale Ltd, its subsidiaries nor their employees accept 
> any responsibility.
>  
> From: Jörn Franke [mailto:jornfra...@gmail.com] 
> Sent: 08 January 2016 06:20
> To: user@hive.apache.org
> Subject: Re: Impact of partitioning on certain queries
>  
> This observation is correct and it is the same  behavior as you see it in 
> other databases supporting partitions. Usually you should avoid many small 
> partitions.
> 
> On 07 Jan 2016, at 23:53, Mich Talebzadeh <m...@peridale.co.uk> wrote:
> 
> Ok we hope that partitioning improves performance where the predicate is on 
> partitioned columns
>  
> I have two tables. One a basic table called smallsales defined as below
>  
> CREATE TABLE `smallsales`(  |
> |   `prod_id` bigint, |
> |   `cust_id` bigint, |
> |   `time_id` timestamp,  |
> |   `channel_id` bigint,  |
> |   `promo_id` bigint,|
> |   `quantity_sold` decimal(10,0),|
> |   `amount_sold` decimal(10,0))  |
> | ROW FORMAT SERDE|
> |   'org.apache.hadoop.hive.serde2.lazy.LazySimpleSerDe'  |
> | STORED

RE: Impact of partitioning on certain queries

[Mich Talebzadeh]

Interesting point below:

 

Well you use a text format for your data so you should not be surprised. For 
text based formats, such as csv, you can always use the hive bitmap index.

 

 

How can one create a bitmap index in Hive please?

 

 

Dr Mich Talebzadeh

 

LinkedIn   
<https://www.linkedin.com/profile/view?id=AAEWh2gBxianrbJd6zP6AcPCCdOABUrV8Pw>
 
https://www.linkedin.com/profile/view?id=AAEWh2gBxianrbJd6zP6AcPCCdOABUrV8Pw

 

Sybase ASE 15 Gold Medal Award 2008

A Winning Strategy: Running the most Critical Financial Data on ASE 15

 
<http://login.sybase.com/files/Product_Overviews/ASE-Winning-Strategy-091908.pdf>
 http://login.sybase.com/files/Product_Overviews/ASE-Winning-Strategy-091908.pdf

Author of the books "A Practitioner’s Guide to Upgrading to Sybase ASE 15", 
ISBN 978-0-9563693-0-7. 

co-author "Sybase Transact SQL Guidelines Best Practices", ISBN 
978-0-9759693-0-4

Publications due shortly:

Complex Event Processing in Heterogeneous Environments, ISBN: 978-0-9563693-3-8

Oracle and Sybase, Concepts and Contrasts, ISBN: 978-0-9563693-1-4, volume one 
out shortly

 

 <http://talebzadehmich.wordpress.com/> http://talebzadehmich.wordpress.com

 

NOTE: The information in this email is proprietary and confidential. This 
message is for the designated recipient only, if you are not the intended 
recipient, you should destroy it immediately. Any information in this message 
shall not be understood as given or endorsed by Peridale Technology Ltd, its 
subsidiaries or their employees, unless expressly so stated. It is the 
responsibility of the recipient to ensure that this email is virus free, 
therefore neither Peridale Ltd, its subsidiaries nor their employees accept any 
responsibility.

 

From: Jörn Franke [mailto:jornfra...@gmail.com] 
Sent: 08 January 2016 08:49
To: user@hive.apache.org
Subject: Re: Impact of partitioning on certain queries

 

Well you use a text format for your data so you should not be surprised. For 
text based formats, such as csv, you can always use the hive bitmap index. I do 
not think it makes a lot of sense to compare here processing csv files and 
internal tables of a relational database.


On 08 Jan 2016, at 09:30, Mich Talebzadeh <m...@peridale.co.uk 
<mailto:m...@peridale.co.uk> > wrote:

Well that is debatable.

 

The following table sales is partitioned in Oracle but has local bitmap indexes 
that help the query.

 

select * from sales where prod_id = 10;

 

no rows selected

 

 

Execution Plan

--

Plan hash value: 511273406

 

-

| Id  | Operation  | Name   | Rows  | Bytes | 
Cost (%CPU)| Time | Pstart| Pstop |

-

|   0 | SELECT STATEMENT   ||   347 | 10063 |   
 93   (0)| 00:00:02 |   |   |

|   1 |  PARTITION RANGE ALL   ||   347 | 10063 |   
 93   (0)| 00:00:02 | 1 |28 |

|   2 |   TABLE ACCESS BY LOCAL INDEX ROWID| SALES  |   347 | 10063 |   
 93   (0)| 00:00:02 | 1 |28 |

|   3 |BITMAP CONVERSION TO ROWIDS ||   |   |   
 |  |   |   |

|*  4 | BITMAP INDEX SINGLE VALUE  | SALES_PROD_BIX |   |   |   
 |  | 1 |28 |

-

 

Obviously at this stage we do not have local indexes in Hive. That could make 
it moredefficient for search and IMO will be a great tool.

 

Cheers,

 

 

Dr Mich Talebzadeh

 

LinkedIn   
<https://www.linkedin.com/profile/view?id=AAEWh2gBxianrbJd6zP6AcPCCdOABUrV8Pw>
 
https://www.linkedin.com/profile/view?id=AAEWh2gBxianrbJd6zP6AcPCCdOABUrV8Pw

 

Sybase ASE 15 Gold Medal Award 2008

A Winning Strategy: Running the most Critical Financial Data on ASE 15

 
<http://login.sybase.com/files/Product_Overviews/ASE-Winning-Strategy-091908.pdf>
 http://login.sybase.com/files/Product_Overviews/ASE-Winning-Strategy-091908.pdf

Author of the books "A Practitioner’s Guide to Upgrading to Sybase ASE 15", 
ISBN 978-0-9563693-0-7. 

co-author "Sybase Transact SQL Guidelines Best Practices", ISBN 
978-0-9759693-0-4

Publications due shortly:

Complex Event Processing in Heterogeneous Environments, ISBN: 978-0-9563693-3-8

Oracle and Sybase, Concepts and Contrasts, ISBN: 978-0-9563693-1-4, volume one 
out shortly

 

 <http://talebzadehmich.wordpress.com/> http://talebzadehmich.wordpress.com

 

NOTE: The information in this email is proprietary and confidential. This 
message is for the designated recipi

RE: Impact of partitioning on certain queries

[Mich Talebzadeh]

Thanks Gopal.

 

Basically the following is true:

 

1.The storage layer is HDFS

2.The execution engine is MR, Tez, Spark etc

3.The access layer is Hive

 

When we say the access layer is Hive, is the assumption correct that we are
referring to optimiser (loosly related to the optimiser in RDBMS). For
example is Hive optimiser aware of the number of underlying partitions. The
reason I am asking this question is that with EXPLAIN I only see Table scan
and it does refer to any partition or partition elimination?

 

 

Cheers

 

 

NOTE: The information in this email is proprietary and confidential. This
message is for the designated recipient only, if you are not the intended
recipient, you should destroy it immediately. Any information in this
message shall not be understood as given or endorsed by Peridale Technology
Ltd, its subsidiaries or their employees, unless expressly so stated. It is
the responsibility of the recipient to ensure that this email is virus free,
therefore neither Peridale Ltd, its subsidiaries nor their employees accept
any responsibility.

 

 

-Original Message-
From: Gopal Vijayaraghavan [mailto:go...@hortonworks.com] On Behalf Of Gopal
Vijayaraghavan
Sent: 08 January 2016 09:34
To: user@hive.apache.org
Subject: Re: Impact of partitioning on certain queries

 

 

> Ok we hope that partitioning improves performance where the predicate 

>is on partitioned columns

 

Nope.

 

Partitioning *only* improves performance if your queries run with

 

set hive.mapred.mode=strict;

 

That's the "use strict" easy way to make sure you're writing good queries.

 

Even then, schema design in hive is something you need to learn with the
assumption that neither the storage layer, nor the compute layer is part of
"hive".

 

It floats itself in an "access" layer above both. Not sure there's any
legacy tech to draw parallels with that.

 

If you haven't seen this before, here's an example of the problem

 

 
<http://www.slideshare.net/Hadoop_Summit/hive-at-yahoo-letters-from-the-tren
>
http://www.slideshare.net/Hadoop_Summit/hive-at-yahoo-letters-from-the-tren

ches/24

 

 

Cheers,

Gopal

Re: Impact of partitioning on certain queries

[Jörn Franke]

https://snippetessay.wordpress.com/2015/07/25/hive-optimizations-with-indexes-bloom-filters-and-statistics/
Maybe a compact index makes more sense if you have high cardinality columns

> On 08 Jan 2016, at 10:11, Mich Talebzadeh <m...@peridale.co.uk> wrote:
> 
> Interesting point below:
>  
> Well you use a text format for your data so you should not be surprised. For 
> text based formats, such as csv, you can always use the hive bitmap index.
>  
>  
> How can one create a bitmap index in Hive please?
>  
>  
> Dr Mich Talebzadeh
>  
> LinkedIn  
> https://www.linkedin.com/profile/view?id=AAEWh2gBxianrbJd6zP6AcPCCdOABUrV8Pw
>  
> Sybase ASE 15 Gold Medal Award 2008
> A Winning Strategy: Running the most Critical Financial Data on ASE 15
> http://login.sybase.com/files/Product_Overviews/ASE-Winning-Strategy-091908.pdf
> Author of the books "A Practitioner’s Guide to Upgrading to Sybase ASE 15", 
> ISBN 978-0-9563693-0-7.
> co-author "Sybase Transact SQL Guidelines Best Practices", ISBN 
> 978-0-9759693-0-4
> Publications due shortly:
> Complex Event Processing in Heterogeneous Environments, ISBN: 
> 978-0-9563693-3-8
> Oracle and Sybase, Concepts and Contrasts, ISBN: 978-0-9563693-1-4, volume 
> one out shortly
>  
> http://talebzadehmich.wordpress.com
>  
> NOTE: The information in this email is proprietary and confidential. This 
> message is for the designated recipient only, if you are not the intended 
> recipient, you should destroy it immediately. Any information in this message 
> shall not be understood as given or endorsed by Peridale Technology Ltd, its 
> subsidiaries or their employees, unless expressly so stated. It is the 
> responsibility of the recipient to ensure that this email is virus free, 
> therefore neither Peridale Ltd, its subsidiaries nor their employees accept 
> any responsibility.
>  
> From: Jörn Franke [mailto:jornfra...@gmail.com] 
> Sent: 08 January 2016 08:49
> To: user@hive.apache.org
> Subject: Re: Impact of partitioning on certain queries
>  
> Well you use a text format for your data so you should not be surprised. For 
> text based formats, such as csv, you can always use the hive bitmap index. I 
> do not think it makes a lot of sense to compare here processing csv files and 
> internal tables of a relational database.
> 
> On 08 Jan 2016, at 09:30, Mich Talebzadeh <m...@peridale.co.uk> wrote:
> 
> Well that is debatable.
>  
> The following table sales is partitioned in Oracle but has local bitmap 
> indexes that help the query.
>  
> select * from sales where prod_id = 10;
>  
> no rows selected
>  
>  
> Execution Plan
> --
> Plan hash value: 511273406
>  
> -
> | Id  | Operation  | Name   | Rows  | Bytes | 
> Cost (%CPU)| Time | Pstart| Pstop |
> -
> |   0 | SELECT STATEMENT   ||   347 | 10063 | 
>93   (0)| 00:00:02 |   |   |
> |   1 |  PARTITION RANGE ALL   ||   347 | 10063 | 
>93   (0)| 00:00:02 | 1 |28 |
> |   2 |   TABLE ACCESS BY LOCAL INDEX ROWID| SALES  |   347 | 10063 | 
>93   (0)| 00:00:02 | 1 |28 |
> |   3 |BITMAP CONVERSION TO ROWIDS ||   |   | 
>|  |   |   |
> |*  4 | BITMAP INDEX SINGLE VALUE  | SALES_PROD_BIX |   |   | 
>|  | 1 |28 |
> -
>  
> Obviously at this stage we do not have local indexes in Hive. That could make 
> it moredefficient for search and IMO will be a great tool.
>  
> Cheers,
>  
>  
> Dr Mich Talebzadeh
>  
> LinkedIn  
> https://www.linkedin.com/profile/view?id=AAEWh2gBxianrbJd6zP6AcPCCdOABUrV8Pw
>  
> Sybase ASE 15 Gold Medal Award 2008
> A Winning Strategy: Running the most Critical Financial Data on ASE 15
> http://login.sybase.com/files/Product_Overviews/ASE-Winning-Strategy-091908.pdf
> Author of the books "A Practitioner’s Guide to Upgrading to Sybase ASE 15", 
> ISBN 978-0-9563693-0-7.
> co-author "Sybase Transact SQL Guidelines Best Practices", ISBN 
> 978-0-9759693-0-4
> Publications due shortly:
> Complex Event Processing in Heterogeneous Environments, ISBN: 
> 978-0-9563693-3-8
> Oracle and Sybase,

Re: Impact of partitioning on certain queries

[Jörn Franke]

This observation is correct and it is the same  behavior as you see it in other 
databases supporting partitions. Usually you should avoid many small partitions.

> On 07 Jan 2016, at 23:53, Mich Talebzadeh  wrote:
> 
> Ok we hope that partitioning improves performance where the predicate is on 
> partitioned columns
>  
> I have two tables. One a basic table called smallsales defined as below
>  
> CREATE TABLE `smallsales`(  |
> |   `prod_id` bigint, |
> |   `cust_id` bigint, |
> |   `time_id` timestamp,  |
> |   `channel_id` bigint,  |
> |   `promo_id` bigint,|
> |   `quantity_sold` decimal(10,0),|
> |   `amount_sold` decimal(10,0))  |
> | ROW FORMAT SERDE|
> |   'org.apache.hadoop.hive.serde2.lazy.LazySimpleSerDe'  |
> | STORED AS INPUTFORMAT   |
> |   'org.apache.hadoop.mapred.TextInputFormat'|
> | OUTPUTFORMAT|
> |   'org.apache.hadoop.hive.ql.io.HiveIgnoreKeyTextOutputFormat'  |
> | LOCATION|
> |   'hdfs://rhes564:9000/user/hive/warehouse/oraclehadoop.db/smallsales'  |
> | TBLPROPERTIES ( |
> |   'COLUMN_STATS_ACCURATE'='true',   |
> |   'last_modified_by'='hduser',  |
> |   'last_modified_time'='1451644705',|
> |   'numFiles'='1',   |
> |   'numRows'='500',  |
> |   'rawDataSize'='193437457',|
> |   'totalSize'='198437457',  |
> |   'transient_lastDdlTime'='1451784743') |
> +-+--+
>  
>  
> So 5 million rows.
>  
>  
> I then created a partitioned table called sales as below
>  
> |createtab_stmt   
>   |
> +---+--+
> | CREATE TABLE `sales`(   
>   |
> |   `prod_id` bigint, 
>   |
> |   `cust_id` bigint, 
>   |
> |   `time_id` timestamp,  
>   |
> |   `channel_id` bigint,  
>   |
> |   `promo_id` bigint,
>   |
> |   `quantity_sold` decimal(10,0),
>   |
> |   `amount_sold` decimal(10,0))  
>   |
> | PARTITIONED BY (
>   |
> |   `year` int,   
>   |
> |   `month` int)  
>   |
> | CLUSTERED BY (  
>   |
> |   prod_id,  
>   |
> |   cust_id,  
>   |
> |   time_id,  
>   |
> |   channel_id,   
>   |
> |   promo_id) 
>   |
> | INTO 256 BUCKETS
>   |
> | ROW FORMAT SERDE
>   |
> |   'org.apache.hadoop.hive.ql.io.orc.OrcSerde'   
>   |
> | STORED AS INPUTFORMAT   
>   |
> |   'org.apache.hadoop.hive.ql.io.orc.OrcInputFormat' 
>   |
> | OUTPUTFORMAT
>   |
> |   'org.apache.hadoop.hive.ql.io.orc.OrcOutputFormat'
>   |
> | LOCATION
>   |
> |   'hdfs://rhes564:9000/user/hive/warehouse/oraclehadoop.db/sales'   
>   |
> | TBLPROPERTIES (