Could someone kindly comment on the below?

Is it allowed (and ok) to use single row functions in the stream definition while using a context partition? (Highlighted in yellow in the EPL below.)

@Name('OutputEvent') create schema OutputEvent as (registration String, status String, ruleName String);

@Name('PerVehicleStream') create context IndividualVehicle partition by registration from VehicleStream;

@Name('FreeParkingSpot') context IndividualVehicle insert into OutputEvent select event2.registration as registration, event2.status as status, "FreeParkingSpot" as ruleName from pattern [every (event1=VehicleStream(hasExited(registration) = false) -> event2=VehicleStream(hasExited(registration) = true))];

I am asking because I get a Null pointer exception while doing the same.

java.lang.NullPointerException

at com.espertech.esper.epl.expression.core.ExprIdentNodeEvaluatorContext.evaluate(ExprIdentNodeEvaluatorContext.java:38)

at com.espertech.esper.epl.expression.dot.ExprDotEvalStaticMethod.get(ExprDotEvalStaticMethod.java:128)

at com.espertech.esper.filter.FilterParamIndexEquals.matchEvent(FilterParamIndexEquals.java:29)

at com.espertech.esper.filter.FilterHandleSetNode.matchEvent(FilterHandleSetNode.java:99)

Hi Esper Dev,

I have a problem where using the between clause with BigDecimals when
selecting from a window.

ExprBetweenCompBigDecimal is doing BigDecimal.equals comparison to test if
the value is equal to the upper or lower bound. BigDecimal.equals checks if
the scale of the two values is equal which we don't want in this case.

Interesting this is not failing when selecting directly from a stream rather
than a window.

The test below should output:

10
15
20
10.000
20.000

But is outputting:

10
15
20

Can't the ExprBetweenComp implementations use the previous result of the
calls to value.compareTo(...) to do the isHighIncluded and isLowIncluded
checks. That would be more efficient too.

Thanks,
From Richard.

package espertest;

import com.espertech.esper.client.*;
import com.tradingscreen.toolbelt.config.ConfigException;

import java.math.BigDecimal;

public class EsperTestBetweenBigDecimal {

public static class Bean {
public BigDecimal decimal;

public Bean(BigDecimal decimal)

public BigDecimal getDecimal()

}

public static void main(String[] args) throws ConfigException

public static class EsperTestListener implements UpdateListener {

@Override
public void update(EventBean[] newEvents, EventBean[] oldEvents) {
if (newEvents != null) {
for (EventBean eventBean : newEvents)

}
}
}
}

In the current version Esper allows access to Non-Relational Data via Method Invocation, see: Polling Method Invocation Results via Iterator

Currently it is only possible to invoke static methods.

This is unfortunately a blocking point for any system using dependency-injection (e.g. Spring and Guice).

I would be great, if method invocations would also be possible on injected objects, e.g.:

select * from AssetMoveEvent, method:myLookupObject.lookupAsset(assetId);

These objects would probably also need to be configured, e.g.:

<lookupObject name="myLookupObject" type="com.espertech.Lookup" />

Then an object could be injected into the EPRuntime, e.g.

runtime.setLookupObject("myLookupObject", obj);

I believe this would be a great enhancement to Esper, finally allowing it to be fully integrated into a dependency injection environment.

As mentioned on the esper-user mailing list, we are regularly getting exceptions from esper. I have managed to get this down to a test case that indicates it is related to a n-ary join.

Notes:

• The test case uses high cpu to provoke the exceptions. Actually running this normally, we don't see high cpu (~5%)
• There are tables mentioned in the join clauses that don't appear in the select clause. In reality we are pulling information out of each of the events but this would just muddy the test case.
• Although it is a complex join, I would expect it to be doing it all inside the context lock for the windows so I wouldn't expect anything else to be modifying the indices at the same time
• It doesn't produce the same exceptions every time but it is generally a mixture of:
  • ConcurrentModificationException
  • ArrayIndexOutOfBoundsException
  • NullPointerException

Validation of EPL (in attach) fails with error below:

Error starting statement: Validation failed in when-not-matched (clause 1): Event type named 'IncidentWindow' has already been declared with differing column name or type information: Type by name 'IncidentWindow' in property 'event' expected event type 'Event' but receives event type 'ChildEvent'

Compiling a contained event selection statement as below generates wrongly formatted EPL code when converting it back to EPL using toEPL() on its EPStatementObjectModel. E.g.:

Module m = engine.getEPAdministrator().getDeploymentAdmin().parse("module some.name; select * from AccountEvent[select * from wallets where currency = 'USD'];");

then iterating through the ModuleItems and compiling them

EPStatementObjectModel compiled = admin.compileEPL(item.getExpression());

then compiled.toEPL() generates the string where the where keyword is missing and a where clauses have no white spaces to the from clause:

"select * from AccountEvent[select * from walletscurrency="USD"]"

Would you consider attached patch that allows Subscribers methods to be named other than "update"?

In projects using dependency injection (e.g. Spring) individual objects (i.e. beans) are often singletons. Up until now it is only possible to use such objects as subscribers if they provide a method named "update".

Attached patch allows to define an arbitrary update method name.

I use following expression to detect address scanning:
and it works well

select src_ip,dst_ip,dst_port,ip from firewall.win:time(30 sec) as fw
group by fw.src_ip "
having count(distinct fw.dst_ip) > 50"
output first every 1 hour"

I decided to use external mysql table scanAllowed consists of ip addresses which is allowd to scan and modified my EPL to use external data
select src_ip,dst_ip,dst_port,ip from firewall.win:time(30 sec) as fw
left outer join sql:mysql ['select ip from scanAllowed'] as allowed on fw.src_ip=allowed.ip
where ip is null
group by fw.src_ip
having count(distinct fw.dst_ip) > 50
output first every 1 hour

I filled table scanAllowed with values 10.0.0.1, 10.0.0.2.
I created test and decided to use caching for sql datasource:

String[] firewallPropsNames =

;
runtime.sendEvent(fwEvent,"firewall");
}
for (int i=0; i<100;i++){
Object [] fwEvent=

;
String[] firewallPropsNames =

;
runtime.sendEvent(fwEvent,"firewall");
}

and results wer

Event received com.espertech.esper.event.map.MapEventType@110ff2b

there is an issue with CurrentTimeEvents when using EsperIO CSV

Please have a look at attached example class and CSV file

The class basically outputs the timestamp associated with each event (CurrentTimeEvent and PriceEvent), in addition the CurrentEPTime is printed.

The result looks like this:
TimeEvent: 0 ep time: 0
Event: 2 ep time: 0
TimeEvent: 2 ep time: 2
Event: 3 ep time: 2
Event: 3 ep time: 2
TimeEvent: 3 ep time: 3
Event: 8 ep time: 3
TimeEvent: 8 ep time: 8

So as one can see the CurrentTimeEvent belonging to the first Event (with timestamp=2) only arrives after the first event. For that reason
CurrentEPTime is still at 2 while the first Event arrives.

I my believe the correct output should look like this:

TimeEvent: 0 ep time: 0
TimeEvent: 2 ep time: 2
Event: 2 ep time: 2
TimeEvent: 3 ep time: 3
Event: 3 ep time: 3
Event: 3 ep time: 3
TimeEvent: 8 ep time: 8
Event: 8 ep time: 8

This time the TimeEvent arrives before the Event and therefore the CurrentEPTime is correct.

I believe attached patch might solve the issue.

OutputValueRenderJSONString.

Characher '>' is escaped using '\>'.
Every json parser fail to render output correctly.

The method is:
com.espertech.esper.event.util.OutputValueRenderJSONString#enquote(String,StringBuilder)

Custom aggregation function with multiple parameters that receives wildcard as one of arguments - for example: foo(*, 100) - receives instance of WildcardParameter class instead of event. This is similar issue to ESPER-662, which was already fixed for single-parameter aggregation functions.

I'm using socket adapter with CSV data format to send events from external system.

I tried to use format:
stream=MyEvent,field1=value without comma,field2="value with, comma",field3=other

but it doesn't work

When I start statement "select * from MyEvent" I see only "value with in field2.

Often under high load the implementation of a collapsing queue (eg: com.lma=
x.collections.coalescing.ring.buffer.CoalescingBuffer) is very useful.
Currently this can be achieved if one disables inbound threading and implements collapsing before submitting to esper (working nicely if external input threads align to context partitions).
However if one wishes to use say hash-segmented contexting then input threading needs to be enabled. This causes all inbound events to placed on a queue and removes the ability to collapse.
Is there anything currently I am unaware of that can help with this (if not I will ponder and raise a feature request).
Raised in jira as my mails to the user group do not succeed.

Recent version of Esper (http://repository.codehaus.org/com/espertech/esper/5.0.0/) has sources.jar attached, but no javadoc.jar. It is needed by documentation tools, such as Dash.

Attaching javadocs is as simple as adding plugin execution into pom.xml:

<plugin>
    <artifactId>maven-javadoc-plugin</artifactId>
    <version>2.9</version>
    <executions>
        <execution>
            <id>attach-javadocs</id>
            <goals>
                <goal>jar</goal>
            </goals>
        </execution>
    </executions>
</plugin>

On 7/8/2014, I submitted a question to the Epser_user Mailing list ( user@esper.codehaus.org) with the subject: " Fwd: EventType Configuration Exception compiling Pojo with Collections." I received no response.

Hence, I am submitting the issue as a bug to the Esper JIRA.

I am seeing an exception I get trying to compile an Esper event Configuration using a Java POJO object. This POJO contains a collection that uses the Java "bounded wildcard" syntax, making it "generic".

I am attaching three sample .java files that illustrate the exception: Document.java, Note.java and PojoEventTest.java (the unit test that drives the scenario).

I am using a POJO event Object in my Esper configuration. The structure of the object hierarchy is a Document may contain many Note(s). Please see the attached .mailing-list-example.zip file (and extract) to refer to the java code that generates the exception.

My EPL statement is:
SELECT d.* FROM document d WHERE d.docNotes.anyOf(x => x.noteId = 7)

The result is the NullPointerException shown in this stack trace:
java.lang.NullPointerException:null [SELECT d.* FROM document d WHERE d.docNotes.anyOf(x => x.noteId = 7)]
at com.espertech.esper.core.service.StatementLifecycleSvcImpl.compile(StatementLifecycleSvcImpl.java:1068)
at com.espertech.esper.core.service.StatementLifecycleSvcImpl.createStopped(StatementLifecycleSvcImpl.java:254)
at com.espertech.esper.core.service.StatementLifecycleSvcImpl.createStoppedAssignName(StatementLifecycleSvcImpl.java:182)
at com.espertech.esper.core.service.StatementLifecycleSvcImpl.createAndStart(StatementLifecycleSvcImpl.java:137)
at com.espertech.esper.core.service.EPAdministratorImpl.createEPLStmt(EPAdministratorImpl.java:117)
at com.espertech.esper.core.service.EPAdministratorImpl.createEPL(EPAdministratorImpl.java:81)
at esper.question.example.PojoEventTest.pojoWithCollectionsEventTest(PojoEventTest.java:30)
at sun.reflect.NativeMethodAccessorImpl.invoke0(Native Method)
at sun.reflect.NativeMethodAccessorImpl.invoke(NativeMethodAccessorImpl.java:57)
at.....

One of the main reasons our team chose Esper as the engine for our Event-processing application was that we could leverage use of existing Java POJO's from our application model. So, we prefer to leverage those existing Java POJOs, if possible.

I do not know if this is expected behavior or a bug. Hoping to get some clarification, direction or get the issue fixed.

The following epl and junit test illustrate issues we are having with queries that join two unique windows receiving high update volumes.
The behaviour only arises when multiple threads are operating, either by sending events in multiple threads with default threading. Or by sending events in a single thread and enabling engine input threading.

Symptoms are:
1.Exceptions of the form.
com.espertech.esper.client.EPException: Unique index violation, index is a unique index and key 'key2' already exists (not sure why this would ocurr on a unique window)
2. Corruptions and incorrect results.

Attached test.
1.Feeds through some events to establish a base state.
2.Fires through many events
3.Feeds through the original events and tests the state.

The unthreaded test case passes and never throws exceptions (this can be broken by enabled engine threading).
The threaded test case throws exceptions and intermittently fails.
Please let me know if anything is unclear.

Code that defines its own aggregation functions (by extending com.espertech.esper.epl.agg.aggregator.AggregationMethod) cannot be used in an OSGi container as the Manifest does not export com.espertech.esper.epl.agg.aggregator package.

I can get round this by embedding the esper dependency in my referencing bundle but this is not ideal

The main uses the wrong provider API, test cases however use the correct one.

Dataflow is sorting the operators to a HashSet class that is not sorted! This prevents the initialization of operators that depend on each other as output channels are not found. To reproduce the problem:

• Create a simple data flow with ~17 operators that depend each other in a chain.
• Run the data flow -> initialization will fail as HashSet will not end up being sorted. Note: this has worked in small test cases as the Java SDK HashSet implementation ends up having the first ~15-16 entries in sorted order (coincident).

Bug is in DataFlowServiceImpl class analyzeBuildOrder function. Note: there is second bug in the sorting loop also - it does not create the order correctly.

Here is a sample fix that seems to work:
private Set<Integer> analyzeBuildOrder(Map<Integer, OperatorDependencyEntry> operators) throws ExprValidationException {

DependencyGraph graph = new DependencyGraph(operators.size(), true);
for (Map.Entry<Integer, OperatorDependencyEntry> entry : operators.entrySet()) {
int myOpNum = entry.getKey();
Set<Integer> incomings = entry.getValue().getIncoming();
for (int incoming : incomings)

}

// Fix 1: use LinkedHashSet
LinkedHashSet<Integer> topDownSet = new LinkedHashSet<Integer>();
while(topDownSet.size() < operators.size()) {

// seconardy sort according to the order of listing
Set<Integer> rootNodes = new TreeSet<Integer>(new Comparator<Integer>() {
public int compare(Integer o1, Integer o2)

});
rootNodes.addAll(graph.getRootNodes(topDownSet));

if (rootNodes.isEmpty()) { // circular dependency could cause this
for (int i = 0; i < operators.size(); i++) {
if (!topDownSet.contains)

}
}

topDownSet.addAll(rootNodes);
}

// Ugly fix 2: invert the output as it was incorrect
LinkedHashSet<Integer> inverted = new LinkedHashSet<Integer>();
Integer[] arr = topDownSet.toArray(new Integer[0]);
for (int i = arr.length - 1; i >= 0; i--)

return inverted;
}

Sample code attached.

If a statement is created with a String containing a natural number as statementName with EPAdministrator.create(EPStatementObjectModel sodaStatement, String statementName) and the statement consists of more than three '->' expressions (meaning A -> B -> C -> D ...), the esper engine crashes with an IndexOutOfBoundsException(index 0, size 0) if the fourth expression (-> D) is processed EVEN if the given statementName wasn't unique and got extended.

Workaround: Use EPAdministrator.createEPL(String, String), where the String should be created via EPStatementObjectModel.toEPL() or use the method mentioned above but with 'real' Strings containing more than natural numbers (e.g. "a9").

This was tested with versions 3.5.0, 4.11.0, 5.0.0.

Method call:

EPStatementObjectModel stmtModel = ... // every A -> B -> C -> D ...
EPAdministrator.create(stmtModel, someString);

...

sendEvent(A);
sendEvent(B);
sendEvent(C);
sendEvent(D); // crashes

The stacktrace is as follows:

ERROR: com.espertech.esper.core.service.ExceptionHandlingService - Exception encountered processing statement '9' statement text 'select * from pattern [every A -> B -> C -> D' : Index: 0, Size: 0
java.lang.IndexOutOfBoundsException: Index: 0, Size: 0
	at java.util.ArrayList.rangeCheck(ArrayList.java:635)
	at java.util.ArrayList.get(ArrayList.java:411)
	at com.espertech.esper.pattern.EvalAndStateNode.generateMatchEvents(EvalAndStateNode.java:260)
	at com.espertech.esper.pattern.EvalAndStateNode.generateMatchEvents(EvalAndStateNode.java:242)
	at com.espertech.esper.pattern.EvalAndStateNode.evaluateTrue(EvalAndStateNode.java:154)
	at com.espertech.esper.pattern.EvalFilterStateNode.evaluateTrue(EvalFilterStateNode.java:80)
	at com.espertech.esper.pattern.EvalFilterStateNode.matchFound(EvalFilterStateNode.java:134)
	at com.espertech.esper.core.service.EPRuntimeImpl.processStatementFilterMultiple(EPRuntimeImpl.java:1212)
	at com.espertech.esper.core.service.EPRuntimeImpl.processMatches(EPRuntimeImpl.java:1054)
	at com.espertech.esper.core.service.EPRuntimeImpl.processWrappedEvent(EPRuntimeImpl.java:465)
	at com.espertech.esper.core.service.EPRuntimeImpl.processEvent(EPRuntimeImpl.java:440)
	at com.espertech.esper.core.service.EPRuntimeImpl.sendEvent(EPRuntimeImpl.java:197)

Hello,

we have problems with an error in error in "esper" for some time now:

java.lang.RuntimeException: com.espertech.esper.client.ConfigurationException: Failed to read schema via URL 'null'
...
at org.apache.catalina.core.StandardContext.listenerStart(StandardContext.java:4939)
at org.apache.catalina.core.StandardContext.startInternal(StandardContext.java:5434)
at org.apache.catalina.util.LifecycleBase.start(LifecycleBase.java:150)
at org.apache.catalina.core.ContainerBase.addChildInternal(ContainerBase.java:901)
at org.apache.catalina.core.ContainerBase.addChild(ContainerBase.java:877)
at org.apache.catalina.core.StandardHost.addChild(StandardHost.java:633)
at org.apache.catalina.startup.HostConfig.deployWAR(HostConfig.java:976)
at org.apache.catalina.startup.HostConfig$DeployWar.run(HostConfig.java:1653)
at java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:471)
at java.util.concurrent.FutureTask.run(FutureTask.java:262)
at java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1145)
at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:615)
at java.lang.Thread.run(Thread.java:724)
Caused by: com.espertech.esper.client.ConfigurationException: Failed to read schema via URL 'null'
at com.espertech.esper.core.service.ConfigurationOperationsImpl.addEventType(ConfigurationOperationsImpl.java:316)
...
Caused by: com.espertech.esper.client.ConfigurationException: Failed to read schema via URL 'null'
at com.espertech.esper.event.xml.XSDSchemaMapper.readSchemaInternal(XSDSchemaMapper.java:131)
at com.espertech.esper.event.xml.XSDSchemaMapper.loadAndMap(XSDSchemaMapper.java:63)
at com.espertech.esper.core.service.ConfigurationOperationsImpl.addEventType(ConfigurationOperationsImpl.java:312)
... 18 more

We found some posts online from people having the same problem. It seemed to have something to do with the java version. The error occurs in java versions since jdk7u40.

Since then we are still running on jdk7_u15 and everything seemed to work.

But lately the error happened even with this version.

We are using:
Esper 4.11.0.
jdk_7_u15

Are there any known issues that can lead to this error except to the java version and are there any known fixes for that problem?

Thank you very much.

Kind regards.

Wolfgang

Queries like 'SELECT MIN(bytes) FROM ...' occurs ASTWalkException.

com/espertech/esper/epl/parse/ASTWalkException.java:27:in `from': com.espertech.esper.epl.parse.ASTWalkException: Uncountered unrecognized min or max node 'MAX'
from com/espertech/esper/epl/parse/ASTLibFunctionHelper.java:302:in `handleMinMax'
from com/espertech/esper/epl/parse/ASTLibFunctionHelper.java:140:in `handleLibFunc'
from com/espertech/esper/epl/parse/EPLTreeWalkerListener.java:302:in `exitLibFunction'
from com/espertech/esper/epl/generated/EsperEPL2GrammarParser.java:13692:in `exitRule'
from org/antlr/v4/runtime/tree/ParseTreeWalker.java:71:in `exitRule'
//snip

It seems that "toLowerCase()" are missing in ASTLibFunctionHelper.java line 302 and 305

We can use compiled regex. Msdn recommends

You should use compiled regular expressions when you call regular expression methods with a specific regular expression relatively frequently.

...

// All Sentences with Interpreted Regex:
// 13,443 matches in 00:00:01.1514100
// All Sentences with Compiled Regex:
// 13,443 matches in 00:00:00.7432921

The way Nesper is supposed to be use is to have limited number of statements and a lot more events. Therefore it's tolerable to spend more time compiling regular expression in a statement once, in favour of better performance gained from not interpreting regex each time we need to evaluate an event.

In our observations with NEsper, compiled regex results in smaller memory footprint and smaller CPU usage when evaluating events.

Bug: Nesper uses String.Format and wraps around a user-passed regex.

From source code (2 occurrences)

_pattern = new Regex(String.Format("^{0}$", patternText));

There are two problems with this:

1) It is an inefficient regex, when processing 100.000 events/sec this may degrade performance for some regex
2) Some regular expressions stop matching

An example for 2) will be: suppose I handle fraudulent transactions. I get stream of suspicious transactions that may or may not be checked. I only want to get a match on the transactions that have not been previously checked. Here is a stream of events

Id | Comment
---------------------
1  | fraud
2  | fraud
3  | [checked] fraud

The regex to match is

(?<!checked\s*)fraud

But because NEsper wraps my regex, the engine receives this

^(?<!checked\s*)fraud$ 

Note the start of the string ^ and end of line $

The last regex will no longer match events I am interested in. This adds much confusion, as some regex will work, whereas the other won't.

Solution: Call to string format should be removed and regex left up to the user to craft.

Nesper un-escapes regular expressions. Suppose I expected the following regex

'\d'

and I want to match a number 123. The bug is that it will never match, as an actual regex that will be used by the engine will be

'd' (notice missing "\")

Sample statement:

select * from Address a
where a.HouseNumber regex '\d'

Sample event

Address.HouseNumber = 123 (no match produced)

When calling the initialize method on EPServiceProvider on a configured engine instance and when the socket io adapter is configured, the initialize causes the configuration to become empty and therefore loosing preconfigured event types.

Example:

Configuration configuration = new Configuration();
configuration.configure("esper.cfg.xml");
EPServiceProvider epService = EPServiceProviderManager.getProvider(engineURI, configuration);
epService.initialize(); // error : EsperIOSocketAdapter overrides configuration
epService.getEPAdministrator().createEPL("select * from DailyPrice");
// get exception if plugin-loader into configuration : com.espertech.esper.client.EPStatementException: Failed to resolve event type: Event type or class named 'DailyPrice' was not found [select * from DailyPrice]

<esper-configuration >
<event-type name="DailyPrice">
<java-util-map>
<map-property name="generic_ticker" class="string"/>
<map-property name="px_open" class="double"/>
<map-property name="px_last" class="double"/>
</java-util-map>
</event-type>
<plugin-loader name="EsperIOSocketAdapter" class-name="com.espertech.esperio.socket.EsperIOSocketAdapterPlugin">
<config-xml>
<esperio-socket-configuration>
<socket name="csvStreamSocket" port="24079" data="csv" />
</esperio-socket-configuration>
</config-xml>
</plugin-loader>
</esper-configuration>

I received the following error when using BigDecimal divide

java.lang.ArithmeticException: Non-terminating decimal expansion; no exact representable decimal result.
at java.math.BigDecimal.divide(Unknown Source)
at com.espertech.esper.type.MathArithTypeEnum$DivideBigDecConvComputer.compute(MathArithTypeEnum.java:721)

This is my statement:
Select sum(price * quantity) / sum(quantity) as avgPrice from Transaction.win:time(1 minute);

When "interval" is specified as uppercase the parser does not allow that.
Workaround is to specify "interval" in lowercase.

Sample EPL:
String s01 = "@Name(\"LogCorrelation\") " +
"SELECT * " +
"FROM LogMessage " +
"MATCH_RECOGNIZE ( " +
" PARTITION BY sessionId " +
" MEASURES S AS start_session, T AS threats, E AS end_session " +
" PATTERN (S T+ E?) " +
" INTERVAL 180 seconds OR terminated " +
" DEFINE " +
" S AS S.type = EsperTestCase4$LogType.TRAFFIC, " +
" T AS T.type = EsperTestCase4$LogType.THREAT AND T.sessionId = S.sessionId, " +
" E AS E.type = EsperTestCase4$LogType.TRAFFIC AND E.sessionId = S.sessionId" +
")";

Sample exception:
Caused by: com.espertech.esper.epl.parse.ASTWalkException: Invalid interval-clause within match-recognize, expecting keyword INTERVAL in text 'INTERVAL 180 seconds OR terminated'

When I register a Map Event type: (http://esper.codehaus.org/esper-4.11.0/doc/reference/en-US/html_single/#eventrep-map-properties)

The Map passed to the configuration with field names and types is not cloned, but referenced directly.
If that Map is cleared or overwriten the original Map fields and types are lost - causing EPL compile errors.

Configuration configuration = new Configuration();
Map<String,Object> typemap = new HashMap<>(1);
typemap.put("val", long.class);
configuration.addEventType("EVENT", typemap);

//Rerun with this line removed
typemap.clear();
EPServiceProvider epService = EPServiceProviderManager.getProvider("MAP Event defn bug", configuration);
EPStatement statement = epService.getEPAdministrator().createEPL(""
+ ""
+ "SELECT val from EVENT "
+ "");

Error: "com.espertech.esper.epl.expression.ExprValidationPropertyException: Property named 'val' is not valid in any stream"

[main] DEBUG c.e.e.c.s.StatementLifecycleSvcImpl - .start Error starting statement
com.espertech.esper.epl.expression.ExprValidationPropertyException: Property named 'val' is not valid in any stream
at com.espertech.esper.epl.expression.ExprIdentNodeUtil.getSuggestionException(ExprIdentNodeUtil.java:210) ~[esper-4.11.0.jar:na]
at com.espertech.esper.epl.expression.ExprIdentNodeUtil.getTypeFromStream(ExprIdentNodeUtil.java:58) ~[esper-4.11.0.jar:na]
at com.espertech.esper.epl.expression.ExprIdentNodeImpl.validate(ExprIdentNodeImpl.java:150) ~[esper-4.11.0.jar:na]
at com.espertech.esper.epl.expression.ExprNodeUtility.getValidatedSubtreeInternal(ExprNodeUtility.java:189) ~[esper-4.11.0.jar:na]
at com.espertech.esper.epl.expression.ExprNodeUtility.getValidatedSubtree(ExprNodeUtility.java:151) ~[esper-4.11.0.jar:na]
at com.espertech.esper.epl.core.ResultSetProcessorFactoryFactory.assignSelectClauseNames(ResultSetProcessorFactoryFactory.java:445) ~[esper-4.11.0.jar:na]
at com.espertech.esper.epl.core.ResultSetProcessorFactoryFactory.getProcessorPrototype(ResultSetProcessorFactoryFactory.java:107) ~[esper-4.11.0.jar:na]
at com.espertech.esper.core.start.EPStatementStartMethodSelectUtil.prepare(EPStatementStartMethodSelectUtil.java:295) ~[esper-4.11.0.jar:na]
at com.espertech.esper.core.start.EPStatementStartMethodSelect.startInternal(EPStatementStartMethodSelect.java:51) ~[esper-4.11.0.jar:na]
at com.espertech.esper.core.start.EPStatementStartMethodBase.start(EPStatementStartMethodBase.java:57) ~[esper-4.11.0.jar:na]
at com.espertech.esper.core.service.StatementLifecycleSvcImpl.startInternal(StatementLifecycleSvcImpl.java:636) [esper-4.11.0.jar:na]
at com.espertech.esper.core.service.StatementLifecycleSvcImpl.start(StatementLifecycleSvcImpl.java:598) [esper-4.11.0.jar:na]
at com.espertech.esper.core.service.StatementLifecycleSvcImpl.createAndStart(StatementLifecycleSvcImpl.java:138) [esper-4.11.0.jar:na]
at com.espertech.esper.core.service.EPAdministratorImpl.createEPLStmt(EPAdministratorImpl.java:117) [esper-4.11.0.jar:na]
at com.espertech.esper.core.service.EPAdministratorImpl.createEPL(EPAdministratorImpl.java:66) [esper-4.11.0.jar:na]
at EsperTest.main(EsperTest.java:31) [classes/:na]
Exception in thread "main" com.espertech.esper.client.EPStatementException: Error starting statement: Property named 'val' is not valid in any stream [SELECT val from EVENT ]
at com.espertech.esper.core.service.StatementLifecycleSvcImpl.startInternal(StatementLifecycleSvcImpl.java:648)
at com.espertech.esper.core.service.StatementLifecycleSvcImpl.start(StatementLifecycleSvcImpl.java:598)
at com.espertech.esper.core.service.StatementLifecycleSvcImpl.createAndStart(StatementLifecycleSvcImpl.java:138)
at com.espertech.esper.core.service.EPAdministratorImpl.createEPLStmt(EPAdministratorImpl.java:117)
at com.espertech.esper.core.service.EPAdministratorImpl.createEPL(EPAdministratorImpl.java:66)
at EsperTest.main(EsperTest.java:31)
Caused by: com.espertech.esper.epl.expression.ExprValidationPropertyException: Property named 'val' is not valid in any stream
at com.espertech.esper.epl.expression.ExprIdentNodeUtil.getSuggestionException(ExprIdentNodeUtil.java:210)
at com.espertech.esper.epl.expression.ExprIdentNodeUtil.getTypeFromStream(ExprIdentNodeUtil.java:58)
at com.espertech.esper.epl.expression.ExprIdentNodeImpl.validate(ExprIdentNodeImpl.java:150)
at com.espertech.esper.epl.expression.ExprNodeUtility.getValidatedSubtreeInternal(ExprNodeUtility.java:189)
at com.espertech.esper.epl.expression.ExprNodeUtility.getValidatedSubtree(ExprNodeUtility.java:151)
at com.espertech.esper.epl.core.ResultSetProcessorFactoryFactory.assignSelectClauseNames(ResultSetProcessorFactoryFactory.java:445)
at com.espertech.esper.epl.core.ResultSetProcessorFactoryFactory.getProcessorPrototype(ResultSetProcessorFactoryFactory.java:107)
at com.espertech.esper.core.start.EPStatementStartMethodSelectUtil.prepare(EPStatementStartMethodSelectUtil.java:295)
at com.espertech.esper.core.start.EPStatementStartMethodSelect.startInternal(EPStatementStartMethodSelect.java:51)
at com.espertech.esper.core.start.EPStatementStartMethodBase.start(EPStatementStartMethodBase.java:57)
at com.espertech.esper.core.service.StatementLifecycleSvcImpl.startInternal(StatementLifecycleSvcImpl.java:636)
... 5 more

(OK, this is a contrived example, but if I have a huge number of event types and fields then for GC reasons it may be beneficial to clear or reuse the Map passed.)

Quick question ...
Using XML configuation how to I define a field in a java.util.Map Event type to contain an array?

e.g.
<event-type name="TestMap">
<java-util-map>
<map-property name="intval" class="int"/>
<map-property name="longarrval" class="long[]"/>
<map-property name="dateval" class="java.util.Date"/>
<map-property name="datearrval" class="java.util.Date[]"/>
</java-util-map>
</event-type>

This fails:
com.espertech.esper.client.ConfigurationException: The type 'java.util.Date[]' is not a recognized type
and a similar error for long[]

I know I can do this programatically, but I'd like to have some mechanism to define the Event as a Map, with Array properties using XML configuration.

For example, can I (programmatically) define a type alias for long[] (e.g. "longarrtype") and then use this in the XML?

<map-property name="longarrval" class="longarrtype"/>
Or can you suggest any other suitable approach?

Example:

expression ABC

context CategorizedContext
select ABC() from MyEvent

The following query works, and returns count_b as expected.
However, if I remove the 'B as b' measure, count_b comes out null, even when B exists.
Test class attached with examples.

select * from Event match_recognize (
measures A as a, B as b, A.uuid as id, B.size() as count_b
pattern (A B*)
interval 1 minute
define
A as (A.evid=1),
B as (B.evid=2)
);

Problem description:

Two statements, with the exact same statement text, are generating different outputs when EngineDefaults.ExecutionConfig.IsPrioritized is set to true and the statements don’t contain an @Priority hint. When @Priority hints (of different priority) are added, the output from the two statements is the same.

The attached file includes the test cases I created:

IsPrioritized_TRUE_Condition_TRUE_PrioritySpecified_TRUE (GOOD)
IsPrioritized_TRUE_Condition_TRUE_PrioritySpecified_FALSE (BAD)
IsPrioritized_TRUE_Condition_FALSE_PrioritySpecified_TRUE (GOOD)
IsPrioritized_TRUE_Condition_FALSE_PrioritySpecified_FALSE (GOOD)

IsPrioritized_FALSE_Condition_TRUE_PrioritySpecified_TRUE (GOOD)
IsPrioritized_FALSE_Condition_TRUE_PrioritySpecified_FALSE (GOOD)
IsPrioritized_FALSE_Condition_FALSE_PrioritySpecified_TRUE (GOOD)
IsPrioritized_FALSE_Condition_FALSE_PrioritySpecified_FALSE (GOOD)

The following is not yet supported:
public void testIt()

public static SupportBean callIt(SupportBean sb)

==========
Workaround is to use "callIt(sb).getIntPrimitive()

I have pre-existing code that I need to integrate with, which defines several classes which implement an interface Identifiable:

public static interface Identifiable<T>

The different classes have different types of identifier and I need to join them to different data. A simple example is attached below.

My problem is that Esper does not use the generic type parameters to compute the type of the "id" property. For example:

Class Node implements Identifiable<String>

Should have an "id" property of type String.

This stackoverflow question and answer provides a solution to resolving the generic type arguments:

http://stackoverflow.com/questions/4876380/generics-and-java-beans-introspector

I have a bean "Node" which is associated with a window and joined to another bean "Data". When I try to invoke the a method on the "Node" with a parameter of type "Data":

public String compute(Data data)

With the query:

select node.id, data.nodeId, data.value, node.compute(data) from NodeDataWindow

I get the following ClassCastException:

ERROR 23Dec 09:27:00.825 [main] ExceptionHandlingService Exception encountered processing statement '60368b01-bd35-446f-b8d9-9da9b5a8647f' statement text 'select node.id, data.nodeId, data.value, node.compute(data) from NodeDataWindow' : com.espertech.esper.event.bean.BeanEventBean cannot be cast to espertest.EsperTest$Node

java.lang.ClassCastException: com.espertech.esper.event.bean.BeanEventBean cannot be cast to espertest.EsperTest$Node

at espertest.EsperTest$Node$$FastClassByCGLIB$$287ce4f2.invoke(<generated>)

at net.sf.cglib.reflect.FastMethod.invoke(FastMethod.java:53)

at com.espertech.esper.epl.expression.ExprDotMethodEvalNoDuck.evaluate(ExprDotMethodEvalNoDuck.java:51)

at com.espertech.esper.epl.expression.ExprDotEvalRootChild.evaluateChain(ExprDotEvalRootChild.java:78)

at com.espertech.esper.epl.expression.ExprDotEvalRootChild.evaluate(ExprDotEvalRootChild.java:51)

at com.espertech.esper.epl.core.eval.EvalSelectNoWildcardMap.process(EvalSelectNoWildcardMap.java:48)

at com.espertech.esper.epl.core.SelectExprResultProcessor.process(SelectExprResultProcessor.java:52)

at com.espertech.esper.epl.core.ResultSetProcessorHandThrough.getSelectEventsNoHaving(ResultSetProcessorHandThrough.java:107)

at com.espertech.esper.epl.core.ResultSetProcessorHandThrough.processViewResult(ResultSetProcessorHandThrough.java:72)

at com.espertech.esper.epl.view.OutputProcessViewDirect.update(OutputProcessViewDirect.java:45)

at com.espertech.esper.view.ViewSupport.updateChildren(ViewSupport.java:115)

at com.espertech.esper.epl.named.NamedWindowConsumerView.update(NamedWindowConsumerView.java:91)

at com.espertech.esper.epl.named.NamedWindowServiceImpl.processHandle(NamedWindowServiceImpl.java:454)

at com.espertech.esper.epl.named.NamedWindowServiceImpl.processDispatches(NamedWindowServiceImpl.java:264)

at com.espertech.esper.epl.named.NamedWindowServiceImpl.dispatch(NamedWindowServiceImpl.java:210)

at com.espertech.esper.core.service.EPRuntimeImpl.processThreadWorkQueueFront(EPRuntimeImpl.java:821)

at com.espertech.esper.core.service.EPRuntimeImpl.processThreadWorkQueue(EPRuntimeImpl.java:773)

at com.espertech.esper.core.service.EPRuntimeImpl.processWrappedEvent(EPRuntimeImpl.java:476)

at com.espertech.esper.core.service.EPRuntimeImpl.processEvent(EPRuntimeImpl.java:437)

at com.espertech.esper.core.service.EPRuntimeImpl.sendEvent(EPRuntimeImpl.java:196)

at espertest.EsperTest.sendEvent(EsperTest.java:99)

at espertest.EsperTest.main(EsperTest.java:83)

I tried making the method take a parameter of type BeanEventBean:

public String compute(BeanEventBean data)

But then I get a type error:

Exception in thread "main" com.espertech.esper.client.EPStatementException: Error starting statement: Failed to resolve method 'compute': Could not find enumeration method, date-time method or instance method named 'compute' in class 'espertest.EsperTest$Node' with matching parameter number and expected parameter type(s) 'Data' (nearest match found was 'compute' taking type(s) 'BeanEventBean') [select node.id, data.nodeId, data.value, node.compute(data) from NodeDataWindow]

at com.espertech.esper.core.service.StatementLifecycleSvcImpl.startInternal(StatementLifecycleSvcImpl.java:637)

at com.espertech.esper.core.service.StatementLifecycleSvcImpl.start(StatementLifecycleSvcImpl.java:595)

at com.espertech.esper.core.service.StatementLifecycleSvcImpl.createAndStart(StatementLifecycleSvcImpl.java:137)

at com.espertech.esper.core.service.EPAdministratorImpl.createEPLStmt(EPAdministratorImpl.java:117)

at com.espertech.esper.core.service.EPAdministratorImpl.createEPL(EPAdministratorImpl.java:66)

at espertest.EsperTest.main(EsperTest.java:79)

at sun.reflect.NativeMethodAccessorImpl.invoke0(Native Method)

at sun.reflect.NativeMethodAccessorImpl.invoke(NativeMethodAccessorImpl.java:57)

at sun.reflect.DelegatingMethodAccessorImpl.invoke(DelegatingMethodAccessorImpl.java:43)

at java.lang.reflect.Method.invoke(Method.java:601)

at com.intellij.rt.execution.application.AppMain.main(AppMain.java:120)

Caused by: com.espertech.esper.epl.expression.ExprValidationException: Failed to resolve method 'compute': Could not find enumeration method, date-time method or instance method named 'compute' in class 'espertest.EsperTest$Node' with matching parameter number and expected parameter type(s) 'Data' (nearest match found was 'compute' taking type(s) 'BeanEventBean')

at com.espertech.esper.epl.expression.ExprDotNodeUtility.getChainEvaluators(ExprDotNodeUtility.java:172)

at com.espertech.esper.epl.expression.ExprDotNode.validate(ExprDotNode.java:290)

at com.espertech.esper.epl.expression.ExprNodeUtility.getValidatedSubtreeInternal(ExprNodeUtility.java:189)

at com.espertech.esper.epl.expression.ExprNodeUtility.getValidatedSubtree(ExprNodeUtility.java:151)

at com.espertech.esper.epl.core.ResultSetProcessorFactoryFactory.assignSelectClauseNames(ResultSetProcessorFactoryFactory.java:450)

at com.espertech.esper.epl.core.ResultSetProcessorFactoryFactory.getProcessorPrototype(ResultSetProcessorFactoryFactory.java:107)

at com.espertech.esper.core.start.EPStatementStartMethodSelectUtil.prepare(EPStatementStartMethodSelectUtil.java:278)

at com.espertech.esper.core.start.EPStatementStartMethodSelect.startInternal(EPStatementStartMethodSelect.java:51)

at com.espertech.esper.core.start.EPStatementStartMethodBase.start(EPStatementStartMethodBase.java:53)

at com.espertech.esper.core.service.StatementLifecycleSvcImpl.startInternal(StatementLifecycleSvcImpl.java:625)

... 10 more

Caused by: com.espertech.esper.epl.expression.ExprValidationException: Failed to resolve method 'compute': Could not find enumeration method, date-time method or instance method named 'compute' in class 'espertest.EsperTest$Node' with matching parameter number and expected parameter type(s) 'Data' (nearest match found was 'compute' taking type(s) 'BeanEventBean')

at com.espertech.esper.epl.expression.ExprDotNodeUtility$2.handle(ExprDotNodeUtility.java:233)

at com.espertech.esper.epl.expression.ExprNodeUtility.resolveMethodAllowWildcardAndStream(ExprNodeUtility.java:637)

at com.espertech.esper.epl.expression.ExprDotNodeUtility.getValidateMethodDescriptor(ExprDotNodeUtility.java:237)

at com.espertech.esper.epl.expression.ExprDotNodeUtility.getChainEvaluators(ExprDotNodeUtility.java:149)

... 19 more

Caused by: com.espertech.esper.epl.core.EngineImportException: Could not find enumeration method, date-time method or instance method named 'compute' in class 'espertest.EsperTest$Node' with matching parameter number and expected parameter type(s) 'Data' (nearest match found was 'compute' taking type(s) 'BeanEventBean')

at com.espertech.esper.epl.core.EngineImportServiceImpl.convert(EngineImportServiceImpl.java:519)

at com.espertech.esper.epl.core.EngineImportServiceImpl.resolveMethod(EngineImportServiceImpl.java:484)

at com.espertech.esper.epl.core.MethodResolutionServiceImpl.resolveMethod(MethodResolutionServiceImpl.java:106)

at com.espertech.esper.epl.expression.ExprNodeUtility.resolveMethodAllowWildcardAndStream(ExprNodeUtility.java:627)

... 21 more

Caused by: com.espertech.esper.epl.core.EngineNoSuchMethodException: Unknown method Node.compute(class espertest.EsperTest$Data)

at com.espertech.esper.util.MethodResolver.resolveMethod(MethodResolver.java:220)

at com.espertech.esper.epl.core.EngineImportServiceImpl.resolveMethod(EngineImportServiceImpl.java:480)

... 23 more

Without the query calling compute the correct data is output, so the join is working correctly.

I’ve tried reducing the complexity of the test but this seems to be the simplest case.

I’ve tried making "compute" a static method taking a Node and Data with similar results.

I’ve tried defining two method one taking "Data" and one taking an "BeanEventBean" but the "Data" on is looked up at query compile time and still used at run time, as expected.

Hi all:
I try to make a stream with a property queried from a named window use following EPL, but get a error

create schema A (symbol string)
create window B.std:lastevent() (a A)
create schema C (a A)
insert into C select (select a from B) @eventbean as a from pattern [timer:interval(1 sec)]

Error starting statement: Event type named 'C' has already been declared with differing column name or type information: Type by name 'C' in property 'a' expected java.lang.String but receives java.util.Map

My questions:
1, property a of B is declared with type A, but why the error message say " 'a' expected java.lang.String"
2, how can i fix this.

Plus:
Because i need to use correlated subquery in future，so i can't use contained-event selection.

Hi Esper users,

I have a stream of "Node" events where each Node has an id and an optional parent id. I’m trying to maintain a window which is a left outer join of a Node with its optional parent Node. The parent event is not necessarily received before the child event. I would like a query that outputs an JoinedNode which includes the node and parent. A event JoinedNode should be produced with no parent if the parent has not been received yet and should update once the parent is received.

The following EPL queries work fine ONLY if I define Node and JoinedNode as POJOs.

create window NodeWindow.std:unique(id) as Node

insert into NodeWindow select * from Node

create window JoinedNodeWindow.std:unique(node.id) as JoinedNode

insert into JoinedNodeWindow select * from JoinedNode

insert into JoinedNode select node, parent from NodeWindow as node left outer join NodeWindow as parent on node.parentId = parent.id

If I try to defined them with create schema:

create schema Node(id string, parentId string, data string)

create schema JoinedNode(node Node, parent Node)

or using "addEventType(String, Map)" I get an error when compiling the "insert into JoinedNode…" statement above:

Caused by: com.espertech.esper.event.EventAdapterException: Event type named 'JoinedNode' has already been declared with differing column name or type information: Type by name 'JoinedNode' in property 'node' expected event type 'Node' but receives event type 'NodeWindow'

I would have thought that the event type NodeWindow would be compatible with Node since NodeWindow is defined "as Node". This seems to be the case when I use POJOs. What am I doing wrong?

I’ve attached my code below. There is a boolean useBeans which can switch between the two modes.

Thanks for any help in advance,

From Richard.

import com.espertech.esper.client.EPAdministrator;

import com.espertech.esper.client.EPRuntime;

import com.espertech.esper.client.EPServiceProvider;

import com.espertech.esper.client.EPServiceProviderManager;

import com.espertech.esper.client.EPStatement;

import com.espertech.esper.client.EventBean;

import com.espertech.esper.client.StatementAwareUpdateListener;

import com.tradingscreen.toolbelt.config.ConfigException;

import java.util.HashMap;

import java.util.Map;

public class EsperTest {

public static class Node {

public String id;

public String parentId;

public String data;

public Node(Map<String, Object> event)

public String getId()

public String getParentId()

public String getData()

@Override

public String toString()

}

public static class JoinedNode {

public Node node;

public Node parent;

public Node getNode()

public void setNode(Node node)

public Node getParent()

public void setParent(Node parent)

}

public static void main(String[] args) throws ConfigException {

EPServiceProvider epService = EPServiceProviderManager.getDefaultProvider();

EPAdministrator epAdministrator = epService.getEPAdministrator();

boolean useBeans = false;

if (useBeans)

else

epAdministrator.createEPL("create window NodeWindow.std:unique(id) as Node");

epAdministrator.createEPL("insert into NodeWindow select * from Node");

addListener(epAdministrator.createEPL("create window JoinedNodeWindow.std:unique(node.id) as JoinedNode"));

epAdministrator.createEPL("insert into JoinedNodeWindow select * from JoinedNode");

epAdministrator.createEPL("insert into JoinedNode select node, parent from NodeWindow as node left outer join NodeWindow as parent on node.parentId = parent.id");

// Node with no parent

System.out.println("Adding 1");

Map<String, Object> event = new HashMap<>();

event.put("id", "1");

sendEvent(epService.getEPRuntime(), event, useBeans);

// Node with already seen parent

System.out.println("Adding 2");

event = new HashMap<>();

event.put("id", "2");

event.put("parentId", "1");

sendEvent(epService.getEPRuntime(), event, useBeans);

// Node with not yet seen parent

System.out.println("Adding 3");

event = new HashMap<>();

event.put("id", "3");

event.put("parentId", "4");

sendEvent(epService.getEPRuntime(), event, useBeans);

// Parent for previous node

System.out.println("Adding 4");

event = new HashMap<>();

event.put("id", "4");

event.put("data", "x");

sendEvent(epService.getEPRuntime(), event, useBeans);

// Join to same parent

System.out.println("Adding 5");

event = new HashMap<>();

event.put("id", "5");

event.put("parentId", "4");

sendEvent(epService.getEPRuntime(), event, useBeans);

// Parent for previous node

System.out.println("Updating 4");

event = new HashMap<>();

event.put("id", "4");

event.put("data", "y");

sendEvent(epService.getEPRuntime(), event, useBeans);

}

private static void sendEvent(EPRuntime epRuntime, Map<String, Object> event, boolean useBeans) {

if (useBeans)

else

}

public static void addListener(EPStatement statement) {

statement.addListener(new StatementAwareUpdateListener() {

@Override

public void update(EventBean[] newEvents, EventBean[] oldEvents, EPStatement statement, EPServiceProvider epServiceProvider)

private void print(EventBean[] events) {

if (events != null) {

for (EventBean eventBean : events) {

System.out.print(" type = " + eventBean.getEventType().getName());

for (String property : eventBean.getEventType().getPropertyNames())

System.out.println();

}

}

}

});

}

}

I faced issue with nested collection in Esper. I have model like this:

myClass3

myClass2

myClass

and try to use query like:

select * from myClass where myClass.class1value.anyOf(x=>x.class2Value.anyOf(y=>y.value like 'string'))

This compiles fine but fails to evaluate correctly.

read up on http://en.wikipedia.org/wiki/Algorithms_for_calculating_variance

Try this:

create schema Test(`order` int);
insert into Test select 0 as `order` from StockTick;

I suggest enumeration methods support tagged-value in pattern as a new kind of input.

We use tagged-value as single value in pattern, and if we can use tagged-value as a collection it will very convenient.

Because sometimes a following event will be matched only if it's properties have some relationship with preceding values(not different tags but same tag with repeated values).

And i think that enumeration method only take values in current run of sub-expression, not entire values in whole lifecycle.

Examples:

Input stream:
1, BarEvent1

5, BarEvent4

6, BarEvent5

7, BarEvent6

8, BarEvent7

9, StopEvent2 {}
10, BarEvent8

11, BarEvent9

Pattern 1:
select b from pattern
[
every ([2] b=BarEvent)
]

Output:
match 1, b=

, c=BarEvent4
match 2: b=

, c=BarEvent8

Pattern 3:
select b,c from pattern
[
every ( (b= BarEvent until StopEvent) -> c= BarEvent(value> b.sumOf(i=>i.value) ) )
]
Output:
match 1, b=

, c=BarEvent4
match 2, b=

, c=BarEvent9

Best regards
zhucj

Sample EPL:
@Priority(1) create context C1 start @now end bar;
@Priority(0) context C1 select symbol,value,context.name,context.id from bar"ï¼›

input events are:

expected output is:
one row output. Because create-context has priority 1 which is higher than the other statement.

actual output is:
context c1

=========== code
epService.getEPAdministrator().getConfiguration().addEventType(MyEvent.class);
epService.getEPAdministrator().createEPL("@Priority(1) create context C1 start @now end MyEvent");
EPStatement stmt = epService.getEPAdministrator().createEPL("@Priority(0) context C1 select symbol,value,context.name,context.id from MyEvent");
stmt.addListener(listener);
epService.getEPRuntime().sendEvent(new MyEvent("S1", 1));
assertEquals(1, listener.getLastNewData().length);

The XSLoader returns null even for the simplest schema.
Older JVM seems to work.

Stack:

Exception in thread "main" com.espertech.esper.client.ConfigurationException: Failed to read schema via URL 'data/AutoIdPmlCore.xsd'
at com.espertech.esper.core.service.EPServicesContextFactoryDefault.init(EPServicesContextFactoryDefault.java:355)
at com.espertech.esper.core.service.EPServicesContextFactoryDefault.createServicesContext(EPServicesContextFactoryDefault.java:90)
at com.espertech.esper.core.service.EPServiceProviderImpl.doInitialize(EPServiceProviderImpl.java:483)
at com.espertech.esper.core.service.EPServiceProviderImpl.<init>(EPServiceProviderImpl.java:89)
at com.espertech.esper.client.EPServiceProviderManager.getProviderInternal(EPServiceProviderManager.java:117)
at com.espertech.esper.client.EPServiceProviderManager.getProvider(EPServiceProviderManager.java:88)
at com.espertech.esper.example.autoid.AutoIdSimMain.run(AutoIdSimMain.java:101)
at com.espertech.esper.example.autoid.AutoIdSimMain.main(AutoIdSimMain.java:73)
Caused by: com.espertech.esper.client.ConfigurationException: Failed to read schema via URL 'data/AutoIdPmlCore.xsd'
at com.espertech.esper.event.xml.XSDSchemaMapper.readSchemaInternal(XSDSchemaMapper.java:130)
at com.espertech.esper.event.xml.XSDSchemaMapper.loadAndMap(XSDSchemaMapper.java:63)
at com.espertech.esper.core.service.EPServicesContextFactoryDefault.init(EPServicesContextFactoryDefault.java:351)
... 7 more

create schema Composite as (x int, y int);
create window A.std:unique(a) as (a string, c Composite);
create schema B as (a string, c Composite);

on B e
merge A w
where e.a = w.a
when matched
then update
set c = e.c;

The above is not allowed. However a function can be called in the set such as "set somefunc(c, e)" as a workaround.

I have a problem when use context with match_recognize.

statements are:
"
create context PerSymbol create context PerSymbol partition by symbol from bar
context PerSymbol select * from bar
match_recognize (
measures R1.symbol as symbol, R1.value as r1_value,R2.value as r2_value
pattern (R1 R2)
define R1 as R1.value >= prev(R1.value),R2 as R2.value >= prev(R2.value)
)"
input events are:

expected output is:

is missing

And this one works fine
"
select * from bar
match_recognize (
partition by symbol
measures R1.symbol as symbol, R1.value as r1_value,R2.value as r2_value
pattern (R1 R2)
define R1 as R1.value >= prev(R1.value),R2 as R2.value >= prev(R2.value)
)"
actual output is:

I believe I have found a defect in Esper 4.10.0.
I try to insert event into named window with fire&forget query as described in 5.15.3 documentation chapter (http://esper.codehaus.org/esper-4.10.0/doc/reference/en-US/html/epl_clauses.html#named_insertfaf). The value in the event I insert is an enum and it seams to be the problem since inserting events with string or primitive values works fine.

See the third line here:

// create schema and window
epServiceProvider.getEPAdministrator().createEPL("CREATE SCHEMA MyMode (mode java.nio.file.AccessMode)");
epServiceProvider.getEPAdministrator().createEPL("CREATE WINDOW CurrentMode.std:unique(mode) MyMode");

// insert
epServiceProvider.getEPRuntime().executeQuery("insert into CurrentMode select java.nio.file.AccessMode.READ as mode");

this make esper to throw NPE:

com.espertech.esper.client.EPStatementException: Error executing statement: null [insert into CurrentAccessMode select java.nio.file.AccessMode.READ as mode]
at com.espertech.esper.core.service.EPRuntimeImpl.getExecuteMethod(EPRuntimeImpl.java:1590)
at com.espertech.esper.core.service.EPRuntimeImpl.executeQueryInternal(EPRuntimeImpl.java:1472)
at com.espertech.esper.core.service.EPRuntimeImpl.executeQuery(EPRuntimeImpl.java:1446)
at com.ocado.ei.monitoring.EventsMonitor.init(EventsMonitor.java:60)
at com.ocado.ei.monitoring.configuration.RulesTest.setUp(RulesTest.java:68)
at sun.reflect.NativeMethodAccessorImpl.invoke0(Native Method)
at sun.reflect.NativeMethodAccessorImpl.invoke(NativeMethodAccessorImpl.java:57)
at sun.reflect.DelegatingMethodAccessorImpl.invoke(DelegatingMethodAccessorImpl.java:43)
at java.lang.reflect.Method.invoke(Method.java:606)
at org.junit.runners.model.FrameworkMethod$1.runReflectiveCall(FrameworkMethod.java:45)
at org.junit.internal.runners.model.ReflectiveCallable.run(ReflectiveCallable.java:15)
at org.junit.runners.model.FrameworkMethod.invokeExplosively(FrameworkMethod.java:42)
at org.junit.internal.runners.statements.RunBefores.evaluate(RunBefores.java:27)
at org.junit.runners.ParentRunner.runLeaf(ParentRunner.java:263)
at org.junit.runners.BlockJUnit4ClassRunner.runChild(BlockJUnit4ClassRunner.java:68)
at org.junit.runners.BlockJUnit4ClassRunner.runChild(BlockJUnit4ClassRunner.java:47)
at org.junit.runners.ParentRunner$3.run(ParentRunner.java:231)
at org.junit.runners.ParentRunner$1.schedule(ParentRunner.java:60)
at org.junit.runners.ParentRunner.runChildren(ParentRunner.java:229)
at org.junit.runners.ParentRunner.access$000(ParentRunner.java:50)
at org.junit.runners.ParentRunner$2.evaluate(ParentRunner.java:222)
at org.junit.runners.ParentRunner.run(ParentRunner.java:300)
at org.mockito.internal.runners.JUnit45AndHigherRunnerImpl.run(JUnit45AndHigherRunnerImpl.java:37)
at org.mockito.runners.MockitoJUnitRunner.run(MockitoJUnitRunner.java:62)
at org.eclipse.jdt.internal.junit4.runner.JUnit4TestReference.run(JUnit4TestReference.java:50)
at org.eclipse.jdt.internal.junit.runner.TestExecution.run(TestExecution.java:38)
at org.eclipse.jdt.internal.junit.runner.RemoteTestRunner.runTests(RemoteTestRunner.java:467)
at org.eclipse.jdt.internal.junit.runner.RemoteTestRunner.runTests(RemoteTestRunner.java:683)
at org.eclipse.jdt.internal.junit.runner.RemoteTestRunner.run(RemoteTestRunner.java:390)
at org.eclipse.jdt.internal.junit.runner.RemoteTestRunner.main(RemoteTestRunner.java:197)
Caused by: java.lang.NullPointerException
at com.espertech.esper.epl.core.SelectExprProcessorHelper.getEvaluator(SelectExprProcessorHelper.java:222)
at com.espertech.esper.epl.core.SelectExprProcessorFactory.getProcessorInternal(SelectExprProcessorFactory.java:179)
at com.espertech.esper.epl.core.SelectExprProcessorFactory.getProcessor(SelectExprProcessorFactory.java:86)
at com.espertech.esper.core.start.EPPreparedExecuteInsertInto.getExecutor(EPPreparedExecuteInsertInto.java:58)
at com.espertech.esper.core.start.EPPreparedExecuteSingleStream.<init>(EPPreparedExecuteSingleStream.java:117)
at com.espertech.esper.core.start.EPPreparedExecuteInsertInto.<init>(EPPreparedExecuteInsertInto.java:32)
at com.espertech.esper.core.service.EPRuntimeImpl.getExecuteMethod(EPRuntimeImpl.java:1567)
... 29 more

However if I replace that line with parametrized query like this:

EPOnDemandPreparedQueryParameterized query = epServiceProvider.getEPRuntime().prepareQueryWithParameters("insert into CurrentMode select ? as mode");

query.setObject(1, java.nio.file.AccessMode.READ);

epServiceProvider.getEPRuntime().executeQuery(query);

It works fine.

As you can see there is some kind of workaround but I would like to add this static insert into statement to my .epl file instead of mixing it with java code.

Sample EPL:
select * from pattern [every [2] (e = Event(p='P1') and not Event(p='P2'))]

===========

I am trying to find pairs of “desired� events that arrive sequentially, without “undesired� events arriving in between.

Events with p=P1 are desired, while events with p=P2 are undesired.

The EPL I am using is:

select * from pattern [every [2] (e = Event(p='P1') and not Event(p='P2'))]

Input events are:

at time t0 = 0 seconds, an event arrives:

event E1

at time t2 = 2 seconds, an event arrives:

event E3

at time t5 = 5 seconds, an event arrives:

event E6

Expected output is:

at time t1 = 1 seconds:

output event

It seems that the pattern stops matching after it reaches undesired event E4

, and so only pair

is fired.

This issue looks similar to http://jira.codehaus.org/browse/ESPER-112, however this was fixed in 1.10 and I’m using version 4.7.0.

Attached is a test example that demonstrates the problem.

When contained-event selection is used in a statement, an instance of com.espertech.esper.epl.property.PropertyEvaluatorAccumulative is created.
Though the instance variable eventsPerStream of PropertyEvaluatorAccumulative is declared as final, eventsPerStream[0] and eventsPerStream[1] (and eventsPerStream[2] , and so on) are set in getAccumulative method and populateEvents method.

When input events are processed in multi-threaded envirionment (e.g. using inbound threading), one instance of PropertyEvaluatorAccumulative is used for a statement, and getAccumulative method and populateEvents method are called from multiple threads. As a result, it is possible that the data of eventsPerStream[0] and eventsPerStream[1] are origined from different input events (and some contained events are missing).

A simple way to fix this problem is to mark the getAccumulative method as synchronized, it may affect the performance.

When a statement starts/stops consuming VDW rstream events, let VDW know. For some virtual data windows, wrapping the events into beans and placing them on the output stream is time-consuming. If the VDW knew that there were no consumers of the window's rstream, the objects could be silently dropped within incurring the performance penalty.

workaround if to wait for the pool to empty before invoking destroy

For group-by aggregation, the hint parameter can already be provided by a variable

Reserved keyword which is acceptable as property name can be used as property name, but when changed some letter(s) to upper case it can't be used as property name in schema-provided XML events.
For example,

select count from SensorEvent

This rule works fine with following event.

<?xml version="1.0" encoding="UTF-8"?>
<Sensor xmlns="SensorSchema">
<count>XXXXXXXX</count>
</Sensor>

But, when some of letter(s) in "count" is changed to upper-case, e.g. "Count", "cOunt", I get following exception.

Exception in thread "main" com.espertech.esper.client.EPStatementException: Error starting statement: Property named 'count' is not valid in any stream (did you mean 'Count'?) [select Count from SensorEvent]
at com.espertech.esper.core.service.StatementLifecycleSvcImpl.startInternal(StatementLifecycleSvcImpl.java:637)
at com.espertech.esper.core.service.StatementLifecycleSvcImpl.start(StatementLifecycleSvcImpl.java:595)
at com.espertech.esper.core.service.StatementLifecycleSvcImpl.createAndStart(StatementLifecycleSvcImpl.java:137)
at com.espertech.esper.core.service.EPAdministratorImpl.createEPLStmt(EPAdministratorImpl.java:117)
at com.espertech.esper.core.service.EPAdministratorImpl.createEPL(EPAdministratorImpl.java:66)
at XMLPropTest.main(XMLPropTest.java:45)
Caused by: com.espertech.esper.epl.expression.ExprValidationPropertyException: Property named 'count' is not valid in any stream (did you mean 'Count'?)
at com.espertech.esper.epl.expression.ExprIdentNodeUtil.getSuggestionException(ExprIdentNodeUtil.java:183)
at com.espertech.esper.epl.expression.ExprIdentNodeUtil.getTypeFromStream(ExprIdentNodeUtil.java:59)
at com.espertech.esper.epl.expression.ExprIdentNodeImpl.validate(ExprIdentNodeImpl.java:150)
at com.espertech.esper.epl.expression.ExprNodeUtility.getValidatedSubtreeInternal(ExprNodeUtility.java:147)
at com.espertech.esper.epl.expression.ExprNodeUtility.getValidatedSubtree(ExprNodeUtility.java:122)
at com.espertech.esper.epl.core.ResultSetProcessorFactoryFactory.getProcessorPrototype(ResultSetProcessorFactoryFactory.java:114)
at com.espertech.esper.core.start.EPStatementStartMethodSelectUtil.prepare(EPStatementStartMethodSelectUtil.java:278)
at com.espertech.esper.core.start.EPStatementStartMethodSelect.startInternal(EPStatementStartMethodSelect.java:51)
at com.espertech.esper.core.start.EPStatementStartMethodBase.start(EPStatementStartMethodBase.java:53)
at com.espertech.esper.core.service.StatementLifecycleSvcImpl.startInternal(StatementLifecycleSvcImpl.java:625)
... 5 more

When I tried "Count" property with no-schema-provided XML event, it does not throw exception but it seems that the property name "Count" is interpreted as "count".

Attached is a sample program which throws the exception above.

During high-volume processing, I have seen stackoverflow errors result when an UpdateListener routes a new event to Esper. The JVM stack size can be increased, but that only delays the issue.

In a simple pattern

every a=A -> (timer:interval(10) and not A)

with @Audit('pattern-instances') and sending As every 1 second, the (not A) pattern keeps increasing while others stay at 1-2.

So I'm not sure if a) the pattern instances are actually increasing or b) @audit is counting erroneously.

This is better expressed in this gist: https://gist.github.com/jsvd/7d1c50042a2be775433a

If the problem is b) it's a minor thing, if a) then it's serious.

For split-stream syntax, it seems that "output all" does not work in combination with prepared statements. Simple test case attached.
Doc link is http://esper.codehaus.org/esper-4.9.0/doc/reference/en-US/html_single/index.html#split_overview