[ 
https://issues.apache.org/jira/browse/JENA-352?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
 ]

Andy Seaborne updated JENA-352:
-------------------------------

    Description: 
The parsers need to keep a bNode label to bNode map which (unusual data) can 
grow too large.  As it takes unusual data, rated as "minor".

outline of solution: 

1/ switch to a bNode allocation scheme which has a seed (a large random number 
per parser run), and concat or XOR with the claimed bNode label to generate a 
unique label without state build up.

2/ (Turtle) don't remember [] bnodes past their usage scope.

3/ Partial - keep a sliding window of bNodes label mappings 

4/ Direct allocation for _:label and tracked allocation for []

e.g.
http://mail-archives.apache.org/mod_mbox/jena-users/201112.mbox/%[email protected]%3E

  was:
The parsers need to keep a bNode label to bNode map which (unusual data) can 
grow too large.  As it takes unusual data, rated as "minor".

outline of solution: 

1/ which to a bNode allocation scheme which is a large random number per file, 
and concat or XOR with the claimed bNode label to generate a unique label 
without state build up.

2/ (Turtle) don't remember [] bnodes past their usage scope.

3/ Partial - keep a sliding window of bNodes label mappings 

4/ Direct allocation for _:label and tracked allocation for []

e.g.
http://mail-archives.apache.org/mod_mbox/jena-users/201112.mbox/%[email protected]%3E

    
> Vast numbers of bNodes can overwhelm the parser
> -----------------------------------------------
>
>                 Key: JENA-352
>                 URL: https://issues.apache.org/jira/browse/JENA-352
>             Project: Apache Jena
>          Issue Type: Bug
>          Components: RIOT, TDB
>            Reporter: Andy Seaborne
>            Priority: Minor
>
> The parsers need to keep a bNode label to bNode map which (unusual data) can 
> grow too large.  As it takes unusual data, rated as "minor".
> outline of solution: 
> 1/ switch to a bNode allocation scheme which has a seed (a large random 
> number per parser run), and concat or XOR with the claimed bNode label to 
> generate a unique label without state build up.
> 2/ (Turtle) don't remember [] bnodes past their usage scope.
> 3/ Partial - keep a sliding window of bNodes label mappings 
> 4/ Direct allocation for _:label and tracked allocation for []
> e.g.
> http://mail-archives.apache.org/mod_mbox/jena-users/201112.mbox/%[email protected]%3E

--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira

Reply via email to