[
https://issues.apache.org/jira/browse/RYA-99?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
David W. Lotts updated RYA-99:
------------------------------
Description:
Update PCJOptimizer's "private static List<ExternalTupleSet>
getAccIndices(Configuration conf)" method to use RyaDetailsRepository to fetch
the names of the PCJs that may be used.
Currently it looks in the configuration file to find all of the PCJ tables that
will be used. If it doesn't find them, then we iterate through the tables and
try to find them programatically. We're moving away from that strategy.
Instead, we should look at the instance's Rya details to find all of the PCJ
IDs that are being maintained, then derive the table names from those. This
means we no longer need to iterate through all the tables to find them using
table operations.
If there are no rya details for the specified instance of Rya, use the old
method.
Here is the file to update:
https://github.com/kchilton2/incubator-rya/blob/RYA-81/extras/indexing/src/main/java/mvm/rya/indexing/pcj/matching/PCJOptimizer.java
The factory for making the tables names is here:
PcjTableNameFactory
was:
Update PCJOptimizer's "private static List<ExternalTupleSet>
getAccIndices(Configuration conf)" method to use RyaDetailsRepository to fetch
the names of the PCJs that may be used.
Currently it looks in the configuration file to find all of the PCJ tables that
will be used. If it doesn't find them, then we iterate through the tables and
try to find them programatically. We're moving away from that strategy.
Instead, we should look at the instance's Rya details to find all of the PCJ
IDs that are being maintained, then derive the table names from those. This
means we no longer need to iterate through all the tables to find them using
table operations.
If there are no rya details for the specified instance of Rya, use the old
method.
Here is the file to update:
https://github.com/kchilton2/incubator-rya/blob/RYA-81/extras/indexing/src/main/java/mvm/rya/indexing/pcj/matching/PCJOptimizer.java
The factory for making the tables names is here:
TODO fill this in.
> Integrate RyaDetails with PCJOptimizer
> --------------------------------------
>
> Key: RYA-99
> URL: https://issues.apache.org/jira/browse/RYA-99
> Project: Rya
> Issue Type: Sub-task
> Affects Versions: 3.2.10
> Reporter: Caleb Meier
> Assignee: David W. Lotts
> Fix For: 3.2.10
>
>
> Update PCJOptimizer's "private static List<ExternalTupleSet>
> getAccIndices(Configuration conf)" method to use RyaDetailsRepository to
> fetch the names of the PCJs that may be used.
> Currently it looks in the configuration file to find all of the PCJ tables
> that will be used. If it doesn't find them, then we iterate through the
> tables and try to find them programatically. We're moving away from that
> strategy. Instead, we should look at the instance's Rya details to find all
> of the PCJ IDs that are being maintained, then derive the table names from
> those. This means we no longer need to iterate through all the tables to find
> them using table operations.
> If there are no rya details for the specified instance of Rya, use the old
> method.
> Here is the file to update:
> https://github.com/kchilton2/incubator-rya/blob/RYA-81/extras/indexing/src/main/java/mvm/rya/indexing/pcj/matching/PCJOptimizer.java
> The factory for making the tables names is here:
> PcjTableNameFactory
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
