[JBoss JIRA] (TEIID-3125) Netezza translator - add support for the LIKE_REGEX predicate.
by Van Halbert (JIRA)
[ https://issues.jboss.org/browse/TEIID-3125?page=com.atlassian.jira.plugin... ]
Van Halbert commented on TEIID-3125:
------------------------------------
What needs to be done to get the rest of the functions supported? Or are you saying, the user must model them as push down to make them work?
> Netezza translator - add support for the LIKE_REGEX predicate.
> --------------------------------------------------------------
>
> Key: TEIID-3125
> URL: https://issues.jboss.org/browse/TEIID-3125
> Project: Teiid
> Issue Type: Enhancement
> Components: Misc. Connectors
> Reporter: Debbie Steigner
> Assignee: Steven Hawkins
> Fix For: 8.9
>
>
> With version 6 of Jboss DV the Netezza translator doesn't support the LIKE_REGEX predicate.
> Netezza doesn't support regular expression by default, but with the installation of SQL Extensions toolkit this functionality is added. After installation Netezza provide the regular expression functions:
> regexp_extract()
> regexp_extract_all()
> regexp_extract_all_sp()
> regexp_extract_sp()
> regexp_instr()
> regexp_like()
> regexp_match_count()
> regexp_replace()
> regexp_replace_sp()
> IBM Netezza SQL Extensions toolkit is an optional package for IBM Netezza data warehouse appliances.
> After installing this toolkit on Netezza and requires that regular expression in JBoss DV syntax are pushed down (delegated) to Netezza to prevent that filtering in the Jboss DV server most be done.
> So please provide a configurable option in the Netezza translator of Jboss DV by which we can enable the LIKE_REGEX predicate for Netezza. This option when enabled will assume that the Netezza SQL extension toolkit is installed and the function regexp_like() is available.
--
This message was sent by Atlassian JIRA
(v6.4.11#64026)
10 years, 7 months
[JBoss JIRA] (TEIID-3693) Add data source support for reading documents (i.e., RTF, DOC, PDF)
by Steven Hawkins (JIRA)
[ https://issues.jboss.org/browse/TEIID-3693?page=com.atlassian.jira.plugin... ]
Steven Hawkins commented on TEIID-3693:
---------------------------------------
Each doc type will need to be a separate issue, unless you are simply indicating extraction at a blob level.
> Add data source support for reading documents (i.e., RTF, DOC, PDF)
> -------------------------------------------------------------------
>
> Key: TEIID-3693
> URL: https://issues.jboss.org/browse/TEIID-3693
> Project: Teiid
> Issue Type: Feature Request
> Components: Misc. Connectors
> Reporter: Van Halbert
> Assignee: Steven Hawkins
>
> Add data source support for reading documents (i.e., RTF, DOC, PDF).
> From a runtime perspective the extraction logic is straight-forward when there is a parsing library. Here's one from CA using JSoup - https://github.com/rokhmanov/teiid-translators/blob/master/translator-scr...
--
This message was sent by Atlassian JIRA
(v6.4.11#64026)
10 years, 7 months
[JBoss JIRA] (TEIID-3652) Teiid could push some outer joins to the source database, but does not
by Steven Hawkins (JIRA)
[ https://issues.jboss.org/browse/TEIID-3652?page=com.atlassian.jira.plugin... ]
Steven Hawkins resolved TEIID-3652.
-----------------------------------
Fix Version/s: 8.12
Resolution: Done
Added RulePlanOuterJoins to look for outer join scenarios that can be rearranged for additional pushdown.
This is currently targeted at pushdown only and for left outer joins. As needed we can expand for full outer joins and for improving the teiid join processing order.
> Teiid could push some outer joins to the source database, but does not
> ----------------------------------------------------------------------
>
> Key: TEIID-3652
> URL: https://issues.jboss.org/browse/TEIID-3652
> Project: Teiid
> Issue Type: Bug
> Components: Query Engine
> Affects Versions: 8.11.2
> Reporter: Mike Higgins
> Assignee: Steven Hawkins
> Fix For: 8.12
>
>
> In the following SQL:
> select *
> from s1.t1 t1
> left outer join s2.t2 t2 on (t1.id = t2.id)
> left outer join s2.t3 t3 on (t2.id2 = t3.id2)
> where t1.name = 'x'
> option makedep s2.t2
>
> where t1 is in one database and t2 and t3 are in a different database,
> Teiid could push the second left outer join to the source database. Currently, it does not and performs the join itself. Some outer joins cannot be pushed, of course, but this one and others like it could be.
--
This message was sent by Atlassian JIRA
(v6.4.11#64026)
10 years, 7 months
[JBoss JIRA] (TEIID-3691) Add ability to set escape tokens within literals
by Steven Hawkins (JIRA)
[ https://issues.jboss.org/browse/TEIID-3691?page=com.atlassian.jira.plugin... ]
Steven Hawkins commented on TEIID-3691:
---------------------------------------
Opened TEIID-3692 to cover string literals in Hive/Impala source sql.
> Add ability to set escape tokens within literals
> ------------------------------------------------
>
> Key: TEIID-3691
> URL: https://issues.jboss.org/browse/TEIID-3691
> Project: Teiid
> Issue Type: Bug
> Components: Query Engine
> Reporter: Don Krapohl
> Assignee: Steven Hawkins
>
> A mismatch of escape sequences exists between Teiid and some databases (Impala for sure). Teiid uses doubled up quotes for single/double whereas some databases expect C-style escapes like \' or \". When string literals are pushed down they fail to parse on the underlying database using Teiid's escapes.
> Teiid uses multiple quotes to escape a single within a string literal. Example would be searching for the string literal Haggley's within single quotes. In Teiid the parser will fail with a parser error using any of these options:
> * Surround with double quotes ("Haggley's")
> * Escape with C-style escape ('Haggley\'s')
> Making the query parse in Teiid the filter value would be 'Haggley''s'. When pushed down into the database this is not valid syntax and should be translated as 'Haggley\'s' or other escape sequence as appropriate to the underlying data store.
--
This message was sent by Atlassian JIRA
(v6.4.11#64026)
10 years, 7 months
[JBoss JIRA] (TEIID-3692) Impala/Hive string literals
by Steven Hawkins (JIRA)
[ https://issues.jboss.org/browse/TEIID-3692?page=com.atlassian.jira.plugin... ]
Steven Hawkins updated TEIID-3692:
----------------------------------
Component/s: Misc. Connectors
(was: Query Engine)
Description: The Hive/Impala logic does not account for the source using c-style strings and will currently for string literals in the source sql use ansi style instead. Bind values are unaffected. (was: A mismatch of escape sequences exists between Teiid and some databases (Impala for sure). Teiid uses doubled up quotes for single/double whereas some databases expect C-style escapes like \' or \". When string literals are pushed down they fail to parse on the underlying database using Teiid's escapes.
Teiid uses multiple quotes to escape a single within a string literal. Example would be searching for the string literal Haggley's within single quotes. In Teiid the parser will fail with a parser error using any of these options:
* Surround with double quotes ("Haggley's")
* Escape with C-style escape ('Haggley\'s')
Making the query parse in Teiid the filter value would be 'Haggley''s'. When pushed down into the database this is not valid syntax and should be translated as 'Haggley\'s' or other escape sequence as appropriate to the underlying data store.)
Fix Version/s: 8.12
Steps to Reproduce: (was: Using the generalized schema-
create table myTable (customerName string)
Issue the following query to any jdbc/odbc database connection with C-type literal quote escapes (such as Hive):
SELECT 1 FROM myTable WHERE customerName = 'Haggley''s'
Expected result is no row (we didn't insert data)
Actual result is a parser error passed back up from the database because it expects single quotes escaped as in 'Haggley\'s')
> Impala/Hive string literals
> ---------------------------
>
> Key: TEIID-3692
> URL: https://issues.jboss.org/browse/TEIID-3692
> Project: Teiid
> Issue Type: Bug
> Components: Misc. Connectors
> Reporter: Don Krapohl
> Assignee: Steven Hawkins
> Fix For: 8.12
>
>
> The Hive/Impala logic does not account for the source using c-style strings and will currently for string literals in the source sql use ansi style instead. Bind values are unaffected.
--
This message was sent by Atlassian JIRA
(v6.4.11#64026)
10 years, 7 months
[JBoss JIRA] (TEIID-3692) Impala/Hive string literals
by Steven Hawkins (JIRA)
Steven Hawkins created TEIID-3692:
-------------------------------------
Summary: Impala/Hive string literals
Key: TEIID-3692
URL: https://issues.jboss.org/browse/TEIID-3692
Project: Teiid
Issue Type: Bug
Components: Query Engine
Reporter: Don Krapohl
Assignee: Steven Hawkins
A mismatch of escape sequences exists between Teiid and some databases (Impala for sure). Teiid uses doubled up quotes for single/double whereas some databases expect C-style escapes like \' or \". When string literals are pushed down they fail to parse on the underlying database using Teiid's escapes.
Teiid uses multiple quotes to escape a single within a string literal. Example would be searching for the string literal Haggley's within single quotes. In Teiid the parser will fail with a parser error using any of these options:
* Surround with double quotes ("Haggley's")
* Escape with C-style escape ('Haggley\'s')
Making the query parse in Teiid the filter value would be 'Haggley''s'. When pushed down into the database this is not valid syntax and should be translated as 'Haggley\'s' or other escape sequence as appropriate to the underlying data store.
--
This message was sent by Atlassian JIRA
(v6.4.11#64026)
10 years, 7 months
[JBoss JIRA] (TEIID-3691) Add ability to set escape tokens within literals
by Don Krapohl (JIRA)
[ https://issues.jboss.org/browse/TEIID-3691?page=com.atlassian.jira.plugin... ]
Don Krapohl resolved TEIID-3691.
--------------------------------
Resolution: Rejected
Will use unescape. Thanks for the response!
> Add ability to set escape tokens within literals
> ------------------------------------------------
>
> Key: TEIID-3691
> URL: https://issues.jboss.org/browse/TEIID-3691
> Project: Teiid
> Issue Type: Bug
> Components: Query Engine
> Reporter: Don Krapohl
> Assignee: Steven Hawkins
>
> A mismatch of escape sequences exists between Teiid and some databases (Impala for sure). Teiid uses doubled up quotes for single/double whereas some databases expect C-style escapes like \' or \". When string literals are pushed down they fail to parse on the underlying database using Teiid's escapes.
> Teiid uses multiple quotes to escape a single within a string literal. Example would be searching for the string literal Haggley's within single quotes. In Teiid the parser will fail with a parser error using any of these options:
> * Surround with double quotes ("Haggley's")
> * Escape with C-style escape ('Haggley\'s')
> Making the query parse in Teiid the filter value would be 'Haggley''s'. When pushed down into the database this is not valid syntax and should be translated as 'Haggley\'s' or other escape sequence as appropriate to the underlying data store.
--
This message was sent by Atlassian JIRA
(v6.4.11#64026)
10 years, 7 months
[JBoss JIRA] (TEIID-3691) Add ability to set escape tokens within literals
by Steven Hawkins (JIRA)
[ https://issues.jboss.org/browse/TEIID-3691?page=com.atlassian.jira.plugin... ]
Steven Hawkins commented on TEIID-3691:
---------------------------------------
If there are sources where string literals are not being translated correctly, then that should be an issue for the translators to be updated. You have identified Impala. It looks like Hive is the same way.
Otherwise, there shouldn't be a broader issue here. Teiid user SQL follows ansi SQL. If you want to have user sql with escaped literals, then there is the unescape Teiid system function.
> Add ability to set escape tokens within literals
> ------------------------------------------------
>
> Key: TEIID-3691
> URL: https://issues.jboss.org/browse/TEIID-3691
> Project: Teiid
> Issue Type: Bug
> Components: Query Engine
> Reporter: Don Krapohl
> Assignee: Steven Hawkins
>
> A mismatch of escape sequences exists between Teiid and some databases (Impala for sure). Teiid uses doubled up quotes for single/double whereas some databases expect C-style escapes like \' or \". When string literals are pushed down they fail to parse on the underlying database using Teiid's escapes.
> Teiid uses multiple quotes to escape a single within a string literal. Example would be searching for the string literal Haggley's within single quotes. In Teiid the parser will fail with a parser error using any of these options:
> * Surround with double quotes ("Haggley's")
> * Escape with C-style escape ('Haggley\'s')
> Making the query parse in Teiid the filter value would be 'Haggley''s'. When pushed down into the database this is not valid syntax and should be translated as 'Haggley\'s' or other escape sequence as appropriate to the underlying data store.
--
This message was sent by Atlassian JIRA
(v6.4.11#64026)
10 years, 7 months