[
https://issues.apache.org/jira/browse/DRILL-3989?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14978979#comment-14978979
]
Julian Hyde commented on DRILL-3989:
------------------------------------
I don’t think Oracle has a clear answer. Their nearest equivalent is v$sql, but
they also have v$process.
JDBC defines the terminology for most people, and they call it statement.
Albeit a JDBC statement can be executed multiple times. I don't know whether
Drill gives each execution a new id, or uses the same statement id for each.
MySQL gets it mixed up: “KILL QUERY terminates the statement the connection is
currently executing, but leaves the connection itself intact.”
https://dev.mysql.com/doc/refman/5.0/en/kill.html
I'd define it as "things that are running that a DBA would like to kill". This
includes SELECT queries, DML and DDL statements. Collectively, statements.
Certainly, INSERT and CREATE TABLE AS SELECT can potentially table as much time
& resources as queries.
> Create a sys.queries table
> --------------------------
>
> Key: DRILL-3989
> URL: https://issues.apache.org/jira/browse/DRILL-3989
> Project: Apache Drill
> Issue Type: Bug
> Components: Metadata
> Reporter: Jacques Nadeau
> Labels: newbie
>
> We should create a sys.queries table that provides a clusterwide view of
> active queries. It could include the following columns:
> queryid, user, sql, current status, number of nodes involved, number of total
> fragments, number of fragments completed, start time
> This should be a pretty straightforward task as we should be able to leverage
> the capabilities around required affinity. A great model to build off of are
> the sys.memory and sys.threads tables.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
