[
https://issues.jboss.org/browse/TEIID-3192?page=com.atlassian.jira.plugin...
]
Steven Hawkins commented on TEIID-3192:
---------------------------------------
Added the initial commit. This keeps our api version at 22, so I moved the force wsc back
to 22 from 26 for consistency. From here we can look at adding additional config
parameters as needed to replace whatever the cxf config could have been used for.
The dynamic metadata does look the same between the old and the new version - except for
some of the native types, which are reported differently such as _boolean rather than
boolean.
I confirmed querying against all tables (although two tables marked as queryable through
the api are actually not at api version 22) and updates. The bulk update logic was
changed to use a single job and add all batches to that job rather than serially
processing - which should improve performance.
I was able to confirm that batching now works and there is a corresponding performance
improvement even for moderately sized results. A sample with 9000 small result rows
through CXF was taking approximately ~8 seconds, vs. ~4.5 seconds through the partner
api.
Migrate Salesforce connectivity to sf partner api jars
------------------------------------------------------
Key: TEIID-3192
URL:
https://issues.jboss.org/browse/TEIID-3192
Project: Teiid
Issue Type: Component Upgrade
Components: Salesforce Connector
Reporter: Steven Hawkins
Assignee: Steven Hawkins
Fix For: 8.10
There are several deficiencies in using CXF for salesforce connectivity - we are not
batching correctly, there is additional CXF overhead, and we have to maintain our own
generated java of the api.
--
This message was sent by Atlassian JIRA
(v6.3.1#6329)