[ 
https://issues.apache.org/jira/browse/IGNITE-21137?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
 ]

Alexander Lapin updated IGNITE-21137:
-------------------------------------
    Description: 
InternalTableImpl's scan methods are currently returning all columns, despite 
of the specified 'columnsToInclude' parameter.
h3. Upd #1

Seems that there are two options:
 # Proper one. Storage scans (e.g. index scan) should support 
'columnsToInclude' for the purposes of late materialization that will be 
extremely useful in case of _some specific_ storage. For such kind of deep 
support, we will probably need to prepare corresponding design.
 # Lightweight one. On PartitionReplicaListener side it's possible to exclude 
all columns that were not mentioned in 'columnsToInclude'. In that case, we 
will reduce the amount of data transferred from replica to client. Seems much 
easier to implement. Has some benefits though.

 

  was:InternalTableImpl's scan methods are currently returning all columns, 
despite of the specified 'columnsToInclude' parameter.


> InternalTableImpl scan methods ignore 'columnsToInclude' parameter
> ------------------------------------------------------------------
>
>                 Key: IGNITE-21137
>                 URL: https://issues.apache.org/jira/browse/IGNITE-21137
>             Project: Ignite
>          Issue Type: Bug
>            Reporter: Ivan Gagarkin
>            Priority: Major
>              Labels: ignite-3
>
> InternalTableImpl's scan methods are currently returning all columns, despite 
> of the specified 'columnsToInclude' parameter.
> h3. Upd #1
> Seems that there are two options:
>  # Proper one. Storage scans (e.g. index scan) should support 
> 'columnsToInclude' for the purposes of late materialization that will be 
> extremely useful in case of _some specific_ storage. For such kind of deep 
> support, we will probably need to prepare corresponding design.
>  # Lightweight one. On PartitionReplicaListener side it's possible to exclude 
> all columns that were not mentioned in 'columnsToInclude'. In that case, we 
> will reduce the amount of data transferred from replica to client. Seems much 
> easier to implement. Has some benefits though.
>  



--
This message was sent by Atlassian Jira
(v8.20.10#820010)

Reply via email to