[
https://issues.apache.org/jira/browse/IGNITE-14906?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Thibaud Faurie resolved IGNITE-14906.
-------------------------------------
Release Note: Missclick : Duplicate of IGNITE-14905
Resolution: Duplicate
> SQL GROUP BY Column not found issue
> -----------------------------------
>
> Key: IGNITE-14906
> URL: https://issues.apache.org/jira/browse/IGNITE-14906
> Project: Ignite
> Issue Type: Bug
> Components: cache, sql, thin client
> Affects Versions: 2.10
> Environment: Ignite server nodes on linux docker containers.
> Application executed in linux docker container.
> Windows 10 pro : DBeaver, VSCode
> My application language : Scala with Apache Ignite Java official library
> Reporter: Thibaud Faurie
> Priority: Major
>
> I have a query that I want to execute through SQL API.
> I have no issue when I run my Query with DBeaver (via Ignite thin client).
> But as soon as I use it in my code with SQL API, it throws an error saying
> that column permission_id is not found.
> More stranger thing, if I pause execution of my code (which generates an
> instance of client node in my cluster) at the SqlFieldsQuery execution step,
> the same error can be thrown by DBeaver sometimes (not 100% accurate)
> However, I figured out this error only occurs when I use the GROUP BY clause
> and some aggregation functions such as GROUP_CONCAT. As soon as I remove
> those, it works like a charm. I suspect some bug with H2 and distributed
> caches.
> Here is my SQL query :
> SELECT app_id, app_label, app_version, app_universal_id, app_status,
> app_manifest_url, app_store_url, app_created_at, app_updated_at,
> CONCAT_WS('||', permission_id, GROUP_CONCAT(content SEPARATOR ';'), code,
> language_id, label, GROUP_CONCAT(text_id SEPARATOR ';')) AS info_data,
> 'PERMISSION_LANG_VARIANT' AS type_data
> FROM (
> SELECT APP.id AS app_id, APP.label AS app_label, PERMISSION.id AS
> permission_id, TEXT.content, LANG.code, TEXT.language_id, LANG.label, TEXT.id
> AS text_id, APP.version AS app_version, APP.app_universal_id AS
> app_universal_id, APP.status AS app_status, APP.manifest_url AS
> app_manifest_url, APP.store_url AS app_store_url, APP.created_at AS
> app_created_at, APP.updated_at AS app_updated_at
> FROM FUSION.APPLICATION AS APP
> INNER JOIN FUSION.PERMISSION AS PERMISSION ON APP.id =
> PERMISSION.application_id
> INNER JOIN FUSION.TEXT AS TEXT ON TEXT.id = PERMISSION.label_text_id
> INNER JOIN FUSION.LANGUAGE AS LANG ON TEXT.language_id = LANG.id
> UNION ALL
> SELECT APP.id AS app_id, APP.label AS app_label, PERMISSION.id AS
> permission_id, TEXT.content, LANG.code, TEXT.language_id, LANG.label, TEXT.id
> AS text_id, APP.version AS app_version, APP.app_universal_id AS
> app_universal_id, APP.status AS app_status, APP.manifest_url AS
> app_manifest_url, APP.store_url AS app_store_url, APP.created_at AS
> app_created_at, APP.updated_at AS app_updated_at
> FROM FUSION.APPLICATION AS APP
> INNER JOIN FUSION.PERMISSION AS PERMISSION ON APP.id =
> PERMISSION.application_id
> INNER JOIN FUSION.TEXT AS TEXT ON TEXT.id = PERMISSION.description_text_id
> INNER JOIN FUSION.LANGUAGE AS LANG ON TEXT.language_id = LANG.id
> )
> GROUP BY permission_id, language_id
> Here is the code I use to execute:
> var igniteQuery = new SqlFieldsQuery(queryString)
> var query = igniteCache.query(igniteQuery)
> query.getAll()
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
