[jira] [Commented] (FLINK-7151) Add a function SQL DDL
[
https://issues.apache.org/jira/browse/FLINK-7151?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16927323#comment-16927323
]
Fabian Hueske commented on FLINK-7151:
--
Hi [~phoenixjiangnan], I agree. Splitting the issue in two is a good approach.
There are a few architectural decisions in the UDF in external JAR feature that
require in my opinion a broader attention.
I don't think we need a full-blown FLIP for the basic syntax, but it would make
sense to start a discussion thread on the dev mailing list to raise the
awareness of others.
Cheers, Fabian
> Add a function SQL DDL
> --
>
> Key: FLINK-7151
> URL: https://issues.apache.org/jira/browse/FLINK-7151
> Project: Flink
> Issue Type: Sub-task
> Components: Table SQL / API
>Reporter: yuemeng
>Assignee: Shuyi Chen
>Priority: Critical
> Fix For: 1.10.0
>
>
> Based on create function and table.we can register a udf,udaf,udtf use sql:
> {code}
> CREATE FUNCTION [IF NOT EXISTS] [catalog_name.db_name.]function_name AS
> class_name [WITH (properties.)];
> DROP FUNCTION [IF EXISTS] [catalog_name.db_name.]function_name;
> ALTER FUNCTION [IF EXISTS] [catalog_name.db_name.]function_name RENAME TO
> new_name;
> {code}
> {code}
> CREATE function 'TOPK' AS
> 'com..aggregate.udaf.distinctUdaf.topk.ITopKUDAF';
> INSERT INTO db_sink SELECT id, TOPK(price, 5, 'DESC') FROM kafka_source GROUP
> BY id;
> {code}
> This ticket can assume that the function class is already loaded in classpath
> by users. Advanced syntax like to how to dynamically load udf libraries from
> external locations can be on a separate ticket.
--
This message was sent by Atlassian Jira
(v8.3.2#803003)
[jira] [Commented] (FLINK-7151) Add a function SQL DDL
[
https://issues.apache.org/jira/browse/FLINK-7151?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16926800#comment-16926800
]
Bowen Li commented on FLINK-7151:
-
Hi [~ZhenqiuHuang] [~fhueske], as mentioned above, I'd prefer to split the
effort, and **have this ticket just deal with basic syntax like the examples
given in the description**, and they are much less controversial. We can assume
that the function class is already in classpath by users.
Advanced syntax like "USING JAR/FILE/ARCHIVE" will need its own FLIP to discuss
questions like those raised by Fabian.
Do you think we need a mini FLIP for just the basic syntax? [~fhueske]
> Add a function SQL DDL
> --
>
> Key: FLINK-7151
> URL: https://issues.apache.org/jira/browse/FLINK-7151
> Project: Flink
> Issue Type: Sub-task
> Components: Table SQL / API
>Reporter: yuemeng
>Assignee: Shuyi Chen
>Priority: Critical
> Fix For: 1.10.0
>
>
> Based on create function and table.we can register a udf,udaf,udtf use sql:
> {code}
> CREATE FUNCTION [IF NOT EXISTS] [catalog_name.db_name.]function_name AS
> class_name [WITH (properties.)];
> DROP FUNCTION [IF EXISTS] [catalog_name.db_name.]function_name;
> ALTER FUNCTION [IF EXISTS] [catalog_name.db_name.]function_name RENAME TO
> new_name;
> {code}
> {code}
> CREATE function 'TOPK' AS
> 'com..aggregate.udaf.distinctUdaf.topk.ITopKUDAF';
> INSERT INTO db_sink SELECT id, TOPK(price, 5, 'DESC') FROM kafka_source GROUP
> BY id;
> {code}
> This ticket can assume that the function class is already loaded in classpath
> by users. Advanced syntax like to how to dynamically load udf libraries from
> external locations can be on a separate ticket.
--
This message was sent by Atlassian Jira
(v8.3.2#803003)
[jira] [Commented] (FLINK-7151) Add a function SQL DDL
[
https://issues.apache.org/jira/browse/FLINK-7151?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16926733#comment-16926733
]
Fabian Hueske commented on FLINK-7151:
--
Hi [~ZhenqiuHuang], thanks for sharing the proposal!
Just a few remarks:
* the proposal assumes that the client side (the one submitting the SQL / Table
API query) can access the external library. It would first be downloaded to the
client and from there shipped to the Flink master and from there distributed
via the blob service
* downloading external code (possibly from a public network) is a security
risk. We should probably add a switch to disable this feature.
* Do we need to think about how to load libraries (JARs) from external catalogs
(like Metastore) or is this is a separate discussion?
Thanks, Fabian
> Add a function SQL DDL
> --
>
> Key: FLINK-7151
> URL: https://issues.apache.org/jira/browse/FLINK-7151
> Project: Flink
> Issue Type: Sub-task
> Components: Table SQL / API
>Reporter: yuemeng
>Assignee: Shuyi Chen
>Priority: Critical
> Fix For: 1.10.0
>
>
> Based on create function and table.we can register a udf,udaf,udtf use sql:
> {code}
> CREATE FUNCTION [IF NOT EXISTS] [catalog_name.db_name.]function_name AS
> class_name [WITH (properties.)];
> DROP FUNCTION [IF EXISTS] [catalog_name.db_name.]function_name;
> ALTER FUNCTION [IF EXISTS] [catalog_name.db_name.]function_name RENAME TO
> new_name;
> {code}
> {code}
> CREATE function 'TOPK' AS
> 'com..aggregate.udaf.distinctUdaf.topk.ITopKUDAF';
> INSERT INTO db_sink SELECT id, TOPK(price, 5, 'DESC') FROM kafka_source GROUP
> BY id;
> {code}
> This ticket can assume that the function class is already loaded in classpath
> by users. Advanced syntax like to how to dynamically load udf libraries from
> external locations can be on a separate ticket.
--
This message was sent by Atlassian Jira
(v8.3.2#803003)
[jira] [Commented] (FLINK-7151) Add a function SQL DDL
[
https://issues.apache.org/jira/browse/FLINK-7151?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16925215#comment-16925215
]
Zhenqiu Huang commented on FLINK-7151:
--
[~phoenixjiangnan] Please try to access it now.
> Add a function SQL DDL
> --
>
> Key: FLINK-7151
> URL: https://issues.apache.org/jira/browse/FLINK-7151
> Project: Flink
> Issue Type: Sub-task
> Components: Table SQL / API
>Reporter: yuemeng
>Assignee: Shuyi Chen
>Priority: Critical
> Fix For: 1.10.0
>
>
> Based on create function and table.we can register a udf,udaf,udtf use sql:
> {code}
> CREATE FUNCTION [IF NOT EXISTS] [catalog_name.db_name.]function_name AS
> class_name [WITH (properties.)];
> DROP FUNCTION [IF EXISTS] [catalog_name.db_name.]function_name;
> ALTER FUNCTION [IF EXISTS] [catalog_name.db_name.]function_name RENAME TO
> new_name;
> {code}
> {code}
> CREATE function 'TOPK' AS
> 'com..aggregate.udaf.distinctUdaf.topk.ITopKUDAF';
> INSERT INTO db_sink SELECT id, TOPK(price, 5, 'DESC') FROM kafka_source GROUP
> BY id;
> {code}
> This ticket can assume that the function class is already loaded in classpath
> by users. Advanced syntax like to how to dynamically load udf libraries from
> external locations can be on a separate ticket.
--
This message was sent by Atlassian Jira
(v8.3.2#803003)
[jira] [Commented] (FLINK-7151) Add a function SQL DDL
[
https://issues.apache.org/jira/browse/FLINK-7151?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16925087#comment-16925087
]
Bowen Li commented on FLINK-7151:
-
[~ZhenqiuHuang] the shared google doc is view-only, can you share one that
people can comment on?
> Add a function SQL DDL
> --
>
> Key: FLINK-7151
> URL: https://issues.apache.org/jira/browse/FLINK-7151
> Project: Flink
> Issue Type: Sub-task
> Components: Table SQL / API
>Reporter: yuemeng
>Assignee: Shuyi Chen
>Priority: Critical
> Fix For: 1.10.0
>
>
> Based on create function and table.we can register a udf,udaf,udtf use sql:
> {code}
> CREATE FUNCTION [IF NOT EXISTS] [catalog_name.db_name.]function_name AS
> class_name [WITH (properties.)];
> DROP FUNCTION [IF EXISTS] [catalog_name.db_name.]function_name;
> ALTER FUNCTION [IF EXISTS] [catalog_name.db_name.]function_name RENAME TO
> new_name;
> {code}
> {code}
> CREATE function 'TOPK' AS
> 'com..aggregate.udaf.distinctUdaf.topk.ITopKUDAF';
> INSERT INTO db_sink SELECT id, TOPK(price, 5, 'DESC') FROM kafka_source GROUP
> BY id;
> {code}
> This ticket can assume that the function class is already loaded in classpath
> by users. Advanced syntax like to how to dynamically load udf libraries from
> external locations can be on a separate ticket.
--
This message was sent by Atlassian Jira
(v8.3.2#803003)
[jira] [Commented] (FLINK-7151) Add a function SQL DDL
[
https://issues.apache.org/jira/browse/FLINK-7151?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16925023#comment-16925023
]
Zhenqiu Huang commented on FLINK-7151:
--
[~phoenixjiangnan][~danny0405] I created a small doc for the implementation.
Please feel free to give any suggestions for improvement.
https://docs.google.com/document/d/1G1pXuxbJ27cQhhayPMKc9EIDbg_M5DTzXlA9nPFfocc
> Add a function SQL DDL
> --
>
> Key: FLINK-7151
> URL: https://issues.apache.org/jira/browse/FLINK-7151
> Project: Flink
> Issue Type: Sub-task
> Components: Table SQL / API
>Reporter: yuemeng
>Assignee: Shuyi Chen
>Priority: Critical
> Fix For: 1.10.0
>
>
> Based on create function and table.we can register a udf,udaf,udtf use sql:
> {code}
> CREATE FUNCTION [IF NOT EXISTS] [catalog_name.db_name.]function_name AS
> class_name [WITH (properties.)];
> DROP FUNCTION [IF EXISTS] [catalog_name.db_name.]function_name;
> ALTER FUNCTION [IF EXISTS] [catalog_name.db_name.]function_name RENAME TO
> new_name;
> {code}
> {code}
> CREATE function 'TOPK' AS
> 'com..aggregate.udaf.distinctUdaf.topk.ITopKUDAF';
> INSERT INTO db_sink SELECT id, TOPK(price, 5, 'DESC') FROM kafka_source GROUP
> BY id;
> {code}
> This ticket can assume that the function class is already loaded in classpath
> by users. Advanced syntax like to how to dynamically load udf libraries from
> external locations can be on a separate ticket.
--
This message was sent by Atlassian Jira
(v8.3.2#803003)
[jira] [Commented] (FLINK-7151) Add a function SQL DDL
[
https://issues.apache.org/jira/browse/FLINK-7151?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16922755#comment-16922755
]
Bowen Li commented on FLINK-7151:
-
Hi [~ZhenqiuHuang], as we discussed offline, we can just focus on enabling very
basic DDL syntax in this ticket. We can assume that the function class is
already loaded in classpath by users. Advanced syntax like to how to
dynamically load udf libraries from external locations can be on a separate
ticket.
> Add a function SQL DDL
> --
>
> Key: FLINK-7151
> URL: https://issues.apache.org/jira/browse/FLINK-7151
> Project: Flink
> Issue Type: Sub-task
> Components: Table SQL / API
>Reporter: yuemeng
>Assignee: Shuyi Chen
>Priority: Critical
> Fix For: 1.10.0
>
>
> Based on create function and table.we can register a udf,udaf,udtf use sql:
> {code}
> CREATE FUNCTION [IF NOT EXISTS] [catalog_name.db_name.]function_name AS
> class_name [WITH (properties.)];
> DROP FUNCTION [IF EXISTS] [catalog_name.db_name.]function_name;
> ALTER FUNCTION [IF EXISTS] [catalog_name.db_name.]function_name RENAME TO
> new_name;
> {code}
> {code}
> CREATE function 'TOPK' AS
> 'com..aggregate.udaf.distinctUdaf.topk.ITopKUDAF';
> INSERT INTO db_sink SELECT id, TOPK(price, 5, 'DESC') FROM kafka_source GROUP
> BY id;
> {code}
> This ticket can assume that the function class is already loaded in classpath
> by users. Advanced syntax like to how to dynamically load udf libraries from
> external locations can be on a separate ticket.
--
This message was sent by Atlassian Jira
(v8.3.2#803003)
[jira] [Commented] (FLINK-7151) Add a function SQL DDL
[
https://issues.apache.org/jira/browse/FLINK-7151?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16920042#comment-16920042
]
Terry Wang commented on FLINK-7151:
---
Hi~ [~ZhenqiuHuang] How is it going on this issue?
Since I am drafting a doc about enhancement for Flink Sql DDL related
Database/Table/Function, I am willing to include this issue and implement it
before 1.10 release. What're your thoughts?
> Add a function SQL DDL
> --
>
> Key: FLINK-7151
> URL: https://issues.apache.org/jira/browse/FLINK-7151
> Project: Flink
> Issue Type: Sub-task
> Components: Table SQL / API
>Reporter: yuemeng
>Assignee: Shuyi Chen
>Priority: Critical
> Fix For: 1.10.0
>
>
> Based on create function and table.we can register a udf,udaf,udtf use sql:
> {code}
> CREATE FUNCTION [IF NOT EXISTS] [catalog_name.db_name.]function_name AS
> class_name;
> DROP FUNCTION [IF EXISTS] [catalog_name.db_name.]function_name;
> ALTER FUNCTION [IF EXISTS] [catalog_name.db_name.]function_name RENAME TO
> new_name;
> {code}
> {code}
> CREATE function 'TOPK' AS
> 'com..aggregate.udaf.distinctUdaf.topk.ITopKUDAF';
> INSERT INTO db_sink SELECT id, TOPK(price, 5, 'DESC') FROM kafka_source GROUP
> BY id;
> {code}
--
This message was sent by Atlassian Jira
(v8.3.2#803003)
[jira] [Commented] (FLINK-7151) Add a function SQL DDL
[
https://issues.apache.org/jira/browse/FLINK-7151?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16914838#comment-16914838
]
Zhenqiu Huang commented on FLINK-7151:
--
[~phoenixjiangnan]Sure. Thanks for the suggestion. I will discuss and draft a
design doc with shuyi early next week.
> Add a function SQL DDL
> --
>
> Key: FLINK-7151
> URL: https://issues.apache.org/jira/browse/FLINK-7151
> Project: Flink
> Issue Type: Sub-task
> Components: Table SQL / API
>Reporter: yuemeng
>Assignee: Shuyi Chen
>Priority: Critical
> Fix For: 1.10.0
>
>
> Based on create function and table.we can register a udf,udaf,udtf use sql:
> {code}
> CREATE FUNCTION [IF NOT EXISTS] [catalog_name.db_name.]function_name AS
> class_name;
> DROP FUNCTION [IF EXISTS] [catalog_name.db_name.]function_name;
> ALTER FUNCTION [IF EXISTS] [catalog_name.db_name.]function_name RENAME TO
> new_name;
> {code}
> {code}
> CREATE function 'TOPK' AS
> 'com..aggregate.udaf.distinctUdaf.topk.ITopKUDAF';
> INSERT INTO db_sink SELECT id, TOPK(price, 5, 'DESC') FROM kafka_source GROUP
> BY id;
> {code}
--
This message was sent by Atlassian Jira
(v8.3.2#803003)
[jira] [Commented] (FLINK-7151) Add a function SQL DDL
[
https://issues.apache.org/jira/browse/FLINK-7151?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16914024#comment-16914024
]
Bowen Li commented on FLINK-7151:
-
[~ZhenqiuHuang] sounds good! please note that 1.10 is targeted to release in
roughly 2 and half months. To make DDL work end-to-end, it may require some
work from catalog API too, thus we'd better start working on this ASAP.
> Add a function SQL DDL
> --
>
> Key: FLINK-7151
> URL: https://issues.apache.org/jira/browse/FLINK-7151
> Project: Flink
> Issue Type: Sub-task
> Components: Table SQL / API
>Reporter: yuemeng
>Assignee: Shuyi Chen
>Priority: Critical
> Fix For: 1.10.0
>
>
> Based on create function and table.we can register a udf,udaf,udtf use sql:
> {code}
> CREATE FUNCTION [IF NOT EXISTS] [catalog_name.db_name.]function_name AS
> class_name;
> DROP FUNCTION [IF EXISTS] [catalog_name.db_name.]function_name;
> ALTER FUNCTION [IF EXISTS] [catalog_name.db_name.]function_name RENAME TO
> new_name;
> {code}
> {code}
> CREATE function 'TOPK' AS
> 'com..aggregate.udaf.distinctUdaf.topk.ITopKUDAF';
> INSERT INTO db_sink SELECT id, TOPK(price, 5, 'DESC') FROM kafka_source GROUP
> BY id;
> {code}
--
This message was sent by Atlassian Jira
(v8.3.2#803003)
[jira] [Commented] (FLINK-7151) Add a function SQL DDL
[
https://issues.apache.org/jira/browse/FLINK-7151?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16913959#comment-16913959
]
Zhenqiu Huang commented on FLINK-7151:
--
[~phoenixjiangnan] As it is already in AthenaX. We would like to contribute
back. I will work with Shuyi on this. Thanks
> Add a function SQL DDL
> --
>
> Key: FLINK-7151
> URL: https://issues.apache.org/jira/browse/FLINK-7151
> Project: Flink
> Issue Type: Sub-task
> Components: Table SQL / API
>Reporter: yuemeng
>Assignee: Shuyi Chen
>Priority: Critical
> Fix For: 1.10.0
>
>
> Based on create temporary function and table.we can register a udf,udaf,udtf
> use sql:
> {code}
> CREATE FUNCTION [IF NOT EXISTS] [catalog_name.db_name.]function_name AS
> class_name;
> DROP FUNCTION [IF EXISTS] [catalog_name.db_name.]function_name;
> ALTER FUNCTION [IF EXISTS] [catalog_name.db_name.]function_name RENAME TO
> new_name;
> {code}
> {code}
> CREATE function 'TOPK' AS
> 'com..aggregate.udaf.distinctUdaf.topk.ITopKUDAF';
> INSERT INTO db_sink SELECT id, TOPK(price, 5, 'DESC') FROM kafka_source GROUP
> BY id;
> {code}
--
This message was sent by Atlassian Jira
(v8.3.2#803003)
[jira] [Commented] (FLINK-7151) Add a function SQL DDL
[
https://issues.apache.org/jira/browse/FLINK-7151?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16911607#comment-16911607
]
Bowen Li commented on FLINK-7151:
-
Hi [~danny0405], can you lead this effort if Shuyi doesn't have time for it?
Thanks!
cc [~ykt836] [~xuefuz]
> Add a function SQL DDL
> --
>
> Key: FLINK-7151
> URL: https://issues.apache.org/jira/browse/FLINK-7151
> Project: Flink
> Issue Type: Sub-task
> Components: Table SQL / API
>Reporter: yuemeng
>Assignee: Shuyi Chen
>Priority: Critical
> Fix For: 1.10.0
>
>
> Based on create temporary function and table.we can register a udf,udaf,udtf
> use sql:
> {code}
> CREATE function [catalog_name.db_name.]function_name AS class_name;
> DROP FUNCTION [IF EXISTS] function_name;
> ALTER FUNCTION [IF EXISTS] function_name RENAME TO new_name;
> {code}
> {code}
> CREATE function 'TOPK' AS
> 'com..aggregate.udaf.distinctUdaf.topk.ITopKUDAF';
> INSERT INTO db_sink SELECT id, TOPK(price, 5, 'DESC') FROM kafka_source GROUP
> BY id;
> {code}
--
This message was sent by Atlassian Jira
(v8.3.2#803003)
[jira] [Commented] (FLINK-7151) Add a function SQL DDL
[
https://issues.apache.org/jira/browse/FLINK-7151?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16910674#comment-16910674
]
Bowen Li commented on FLINK-7151:
-
Hi [~suez1224] , are you still interested in moving this forward?
BTW, shall we differentiate DDL for function and DDL for temp function? I think
they probably should be two JIRAs
> Add a function SQL DDL
> --
>
> Key: FLINK-7151
> URL: https://issues.apache.org/jira/browse/FLINK-7151
> Project: Flink
> Issue Type: Sub-task
> Components: Table SQL / API
>Reporter: yuemeng
>Assignee: Shuyi Chen
>Priority: Major
>
> Based on create temporary function and table.we can register a udf,udaf,udtf
> use sql:
> {code}
> CREATE TEMPORARY function 'TOPK' AS
> 'com..aggregate.udaf.distinctUdaf.topk.ITopKUDAF';
> INSERT INTO db_sink SELECT id, TOPK(price, 5, 'DESC') FROM kafka_source GROUP
> BY id;
> {code}
--
This message was sent by Atlassian Jira
(v8.3.2#803003)
