[JBoss JIRA] (TEIID-147) Support for JAVA Stored Procedures
by Steven Hawkins (JIRA)
[ https://issues.jboss.org/browse/TEIID-147?page=com.atlassian.jira.plugin.... ]
Steven Hawkins commented on TEIID-147:
--------------------------------------
> Each factory could return an array/iterable of stored procedures, each of which implemented some interface.
The further we get down that path, the more things start to look like our typical connector api - albeit with a little more ease of use.
Have a look at the handling for scalar functions - see MetadataFactory.addFunction(String, Class) and the internal org.teiid.query.function.TeiidFunction annotation that we use for some of the system functions. Some amount of that convention / logic could be added for procedures as well.
> The second option seems like it is better suited to stored procedures are better authored in a text file.
The though here is more about Teiid Designer users who want some additional language flexibility without getting too much into things like modules, classloaders, etc. However as you point out without good IDE support it's unlikely that you'd actually want to use this path.
> Support for JAVA Stored Procedures
> ----------------------------------
>
> Key: TEIID-147
> URL: https://issues.jboss.org/browse/TEIID-147
> Project: Teiid
> Issue Type: Feature Request
> Security Level: Public(Everyone can see)
> Components: Query Engine
> Affects Versions: 9.x
> Reporter: Ramesh Reddy
> Fix For: 9.0
>
>
> Currently the Stored procedure support is lacking for both our procedure language and XQuery. Procedure support could be extended to java (or groovy) code, which will provide a powerful engine to execute a arbitrary queries.
--
This message was sent by Atlassian JIRA
(v6.2.6#6264)
11 years, 9 months
[JBoss JIRA] (TEIID-147) Support for JAVA Stored Procedures
by Mark Addleman (JIRA)
[ https://issues.jboss.org/browse/TEIID-147?page=com.atlassian.jira.plugin.... ]
Mark Addleman commented on TEIID-147:
-------------------------------------
On the first option, I'd prefer a convention-based approach rather than annotations because annotation support in Clojure is still experimental. I don't know how well other languages support annotations. Spit-balling an idea: to better support dynamically compiled languages, clients could register one or more stored-proc factories. Each factory could return an array/iterable of stored procedures, each of which implemented some interface.
It seems like these two options address two different use cases. The first option addresses stored procedures authored in a language-aware IDE. The second option seems like it is better suited to stored procedures are better authored in a text file. As I see it, the primary difference between the two is whether the IDE could understand the stored procedure as a language or simply as text. Of course, if the IDE understands the stored procedure as a real language, the programmer has access to all of the IDE niceties.
My vote is for the first option.
> Support for JAVA Stored Procedures
> ----------------------------------
>
> Key: TEIID-147
> URL: https://issues.jboss.org/browse/TEIID-147
> Project: Teiid
> Issue Type: Feature Request
> Security Level: Public(Everyone can see)
> Components: Query Engine
> Affects Versions: 9.x
> Reporter: Ramesh Reddy
> Fix For: 9.0
>
>
> Currently the Stored procedure support is lacking for both our procedure language and XQuery. Procedure support could be extended to java (or groovy) code, which will provide a powerful engine to execute a arbitrary queries.
--
This message was sent by Atlassian JIRA
(v6.2.6#6264)
11 years, 9 months
[JBoss JIRA] (TEIID-147) Support for JAVA Stored Procedures
by Steven Hawkins (JIRA)
[ https://issues.jboss.org/browse/TEIID-147?page=com.atlassian.jira.plugin.... ]
Steven Hawkins commented on TEIID-147:
--------------------------------------
There are a couple of approaches here. Given our expanded metadata capabilities, the first approach is conventions/annotations to map a java class/method to a procedure invocation - I believe derby does something similar.
The other which is more what this was originally thinking of would be similar to pg and other dbs that support alternative languages inline with the ddl such that there are conventions for how inputs/outputs and result sets are bound and returned. On this latter one using the pluggable script engine infrastructure (just like the object_table construct) you can utilize a variety of vm languages. The only downside is that the execution of the procedure becomes thread bound.
> Support for JAVA Stored Procedures
> ----------------------------------
>
> Key: TEIID-147
> URL: https://issues.jboss.org/browse/TEIID-147
> Project: Teiid
> Issue Type: Feature Request
> Security Level: Public(Everyone can see)
> Components: Query Engine
> Affects Versions: 9.x
> Reporter: Ramesh Reddy
> Fix For: 9.0
>
>
> Currently the Stored procedure support is lacking for both our procedure language and XQuery. Procedure support could be extended to java (or groovy) code, which will provide a powerful engine to execute a arbitrary queries.
--
This message was sent by Atlassian JIRA
(v6.2.6#6264)
11 years, 9 months
[JBoss JIRA] (TEIID-3023) records lost when query by hiernate
by Steven Hawkins (JIRA)
[ https://issues.jboss.org/browse/TEIID-3023?page=com.atlassian.jira.plugin... ]
Steven Hawkins commented on TEIID-3023:
---------------------------------------
You'll need to provide the query plan to see what is being processed and what the source queries are.
> records lost when query by hiernate
> -----------------------------------
>
> Key: TEIID-3023
> URL: https://issues.jboss.org/browse/TEIID-3023
> Project: Teiid
> Issue Type: Bug
> Security Level: Public(Everyone can see)
> Affects Versions: 7.3
> Environment: window7,eclipse Helios Service Release 2, Teiid v7.3.0
> Reporter: Lai Chao
> Priority: Minor
>
> Recently, I get an issue when retrieve data from vdb that some records lost in the list. Fox example, from source table we can get 12100 records, but from vdb query(projects with hibernate) can only get 10800+ records.
> Meanwhile, there's a set on maxSetSize = 1000, is there any affects?
> Please advise.
--
This message was sent by Atlassian JIRA
(v6.2.6#6264)
11 years, 9 months
[JBoss JIRA] (TEIID-147) Support for JAVA Stored Procedures
by Mark Addleman (JIRA)
[ https://issues.jboss.org/browse/TEIID-147?page=com.atlassian.jira.plugin.... ]
Mark Addleman commented on TEIID-147:
-------------------------------------
I'm unsure of the implementation you're thinking but I'd favor one that made it easy to supply any JVM-based language as a stored procedure, particularly Clojure
> Support for JAVA Stored Procedures
> ----------------------------------
>
> Key: TEIID-147
> URL: https://issues.jboss.org/browse/TEIID-147
> Project: Teiid
> Issue Type: Feature Request
> Security Level: Public(Everyone can see)
> Components: Query Engine
> Affects Versions: 9.x
> Reporter: Ramesh Reddy
> Fix For: 9.0
>
>
> Currently the Stored procedure support is lacking for both our procedure language and XQuery. Procedure support could be extended to java (or groovy) code, which will provide a powerful engine to execute a arbitrary queries.
--
This message was sent by Atlassian JIRA
(v6.2.6#6264)
11 years, 9 months
[JBoss JIRA] (TEIID-3023) records lost when query by hiernate
by Lai Chao (JIRA)
[ https://issues.jboss.org/browse/TEIID-3023?page=com.atlassian.jira.plugin... ]
Lai Chao commented on TEIID-3023:
---------------------------------
Hi, finnally I get access to the vdb by eclipse. But the situation is weird.
I tried to select from 3 tables, it also get 10800+ records with some data miss.
But I can get records if I add the lost data in condition.
for example the sql 1 can get 10800+records, with data miss "centre='10001'".
And I add the condition "centre='10001'" in the script as sql 2, I can get the data.
Please advise.
sql 1:
select distinct a.columns,b.columns,c.columns
from table_a a, table_b b,table_c c
where a.centre = b.code and b.code = c.division
and a.org='A00';
sql 2:
select distinct a.columns,b.columns,c.columns
from table_a a, table_b b,table_c c
where a.centre = b.code and b.code = c.division
and a.org='A00'
and centre = '10001';
> records lost when query by hiernate
> -----------------------------------
>
> Key: TEIID-3023
> URL: https://issues.jboss.org/browse/TEIID-3023
> Project: Teiid
> Issue Type: Bug
> Security Level: Public(Everyone can see)
> Affects Versions: 7.3
> Environment: window7,eclipse Helios Service Release 2, Teiid v7.3.0
> Reporter: Lai Chao
> Priority: Minor
>
> Recently, I get an issue when retrieve data from vdb that some records lost in the list. Fox example, from source table we can get 12100 records, but from vdb query(projects with hibernate) can only get 10800+ records.
> Meanwhile, there's a set on maxSetSize = 1000, is there any affects?
> Please advise.
--
This message was sent by Atlassian JIRA
(v6.2.6#6264)
11 years, 9 months
[JBoss JIRA] (TEIID-3029) Duplicate classes with bundle builds
by Steven Hawkins (JIRA)
[ https://issues.jboss.org/browse/TEIID-3029?page=com.atlassian.jira.plugin... ]
Steven Hawkins commented on TEIID-3029:
---------------------------------------
It's not an issue with provided, it's that the exported-classes includes all teiid packages with causes the inclusion of the other jars.
> Duplicate classes with bundle builds
> ------------------------------------
>
> Key: TEIID-3029
> URL: https://issues.jboss.org/browse/TEIID-3029
> Project: Teiid
> Issue Type: Bug
> Security Level: Public(Everyone can see)
> Affects Versions: 8.6
> Reporter: Steven Hawkins
> Assignee: Steven Hawkins
> Fix For: 8.7.1, 8.8
>
>
> An issue with TEIID-2424 is that teiid internal dependencies marked as provided are added to the bundle jar, which results in duplicate classes.
--
This message was sent by Atlassian JIRA
(v6.2.6#6264)
11 years, 9 months