[
https://issues.apache.org/jira/browse/FLINK-18443?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17147494#comment-17147494
]
mzz commented on FLINK-18443:
-----------------------------
I tried to read the source code，But it didn't help me。Because of the parameter
METRICS_OPERATOR_NAME_MAX_LENGTH is final。
{code:java}
static final int METRICS_OPERATOR_NAME_MAX_LENGTH = 80;
public OperatorMetricGroup getOrAddOperator(OperatorID operatorID, String name)
{
if (name != null && name.length() >
METRICS_OPERATOR_NAME_MAX_LENGTH) {
LOG.warn("The operator name {} exceeded the {}
characters length limit and was truncated.", name,
METRICS_OPERATOR_NAME_MAX_LENGTH);
name = name.substring(0,
METRICS_OPERATOR_NAME_MAX_LENGTH);
}
OperatorMetricGroup operator = new
OperatorMetricGroup(this.registry, this, operatorID, name);
// unique OperatorIDs only exist in streaming, so we have to
rely on the name for batch operators
final String key = operatorID + name;
synchronized (this) {
OperatorMetricGroup previous = operators.put(key,
operator);
if (previous == null) {
// no operator group so far
return operator;
} else {
// already had an operator group. restore that
one.
operators.put(key, previous);
return previous;
}
}
}
{code}
> The operator name select: (ip, ts, count, environment.access AS access,
> environment.brand AS brand, sid, params.adid AS adid, eventid) exceeded the
> 80 characters length limit and was truncated
> ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
>
> Key: FLINK-18443
> URL: https://issues.apache.org/jira/browse/FLINK-18443
> Project: Flink
> Issue Type: Bug
> Components: API / DataStream
> Affects Versions: 1.9.0
> Reporter: mzz
> Priority: Major
>
> *Schema:*
> {code:java}
> .withSchema(new Schema()
> .field("ip", Types.STRING())
> .field("ts", Types.STRING())
> .field("procTime", Types.SQL_TIMESTAMP).proctime()
> .field("environment", schemaEnvironment)
> .field("advs", ObjectArrayTypeInfo.getInfoFor(new Array[Row](0).getClass,
> Types.ROW(Array("count","sid", "eventid","params"),
> Array[TypeInformation[_]](Types.STRING(),Types.STRING(),
>
> Types.STRING(),Types.ROW(Array("adid","adtype","ecpm"),Array[TypeInformation[_]]
> (Types.STRING(),Types.STRING(),Types.STRING()))))))
> )
> {code}
> *when execute this sql*:
> {code:java}
> val sql =
> """
> |SELECT
> |ip,
> |ts,
> |params.ad,
> |params.adtype,
> |eventid,
> |procTime
> |FROM aggs_test
> |CROSS JOIN UNNEST(advs) AS t (`count`,sid, eventid,params)
> |""".stripMargin
> {code}
> *I got a warning，and the console keeps brushing this warning,no normal
> printout*
> {code:java}
> 09:38:38,694 WARN org.apache.flink.metrics.MetricGroup
> - The operator name correlate: table(explode($cor0.advs)), select: ip, ts,
> procTime, advs, sid, eventid, params exceeded the 80 characters length limit
> and was truncated.
> {code}
> *But after I change it to this way, although I occasionally brush this Warn,
> it can be output normally。I change the 'params' type from Types.ROW to
> Types.STRING*。
> {code:java}
> .field("advs", ObjectArrayTypeInfo.getInfoFor(new Array[Row](0).getClass,
> Types.ROW(Array("count", "sid", "eventid", "params"),
> Array[TypeInformation[_]](Types.STRING(), Types.STRING(),
> Types.STRING(), Types.STRING()))))
> {code}
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
