You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@flink.apache.org by "Timo Walther (JIRA)" <ji...@apache.org> on 2016/02/24 13:55:18 UTC
[jira] [Updated] (FLINK-3497) Add SQL scalar functions to Table API
[ https://issues.apache.org/jira/browse/FLINK-3497?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Timo Walther updated FLINK-3497:
--------------------------------
Description:
In order to use the Table API as ETL tool and prepare for FLINK-2099, we need to add more scalar functions such as {{trim()}}, {{abs()}}, {{like()}}, etc.
Calcite implements the most important functions. We can basically forward calls to Calcites built-in runtime functions. Some functions need special treatment because of Flink specifics.
I would propose the following steps:
- Implement TRIM, SUBSTRING as reference design
- remaining string functions
- math functions
- Date/time functions
- System functions
- Case function
- Array functions
- other
Each step includes implementation, test and documentation.
was:
In order to use the Table API as ETL tool and prepare for FLINK-2099, we need to add more scalar functions such as {{trim()}}, {{abs()}}, {{like()}}, etc.
Calcite implements the most important functions. We can basically forward calls to Calcites built-in runtime functions. Some functions need special treatment because of Flink specifics.
I would propose the following steps:
- Implement TRIM, SUBSTRING as reference design
- remaining string functions
- math functions
- Date/time functions
- System functions
- Case
- Array
- other
Each step includes implementation, test and documentation.
> Add SQL scalar functions to Table API
> -------------------------------------
>
> Key: FLINK-3497
> URL: https://issues.apache.org/jira/browse/FLINK-3497
> Project: Flink
> Issue Type: New Feature
> Components: Table API
> Reporter: Timo Walther
>
> In order to use the Table API as ETL tool and prepare for FLINK-2099, we need to add more scalar functions such as {{trim()}}, {{abs()}}, {{like()}}, etc.
> Calcite implements the most important functions. We can basically forward calls to Calcites built-in runtime functions. Some functions need special treatment because of Flink specifics.
> I would propose the following steps:
> - Implement TRIM, SUBSTRING as reference design
> - remaining string functions
> - math functions
> - Date/time functions
> - System functions
> - Case function
> - Array functions
> - other
> Each step includes implementation, test and documentation.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)