Hi Denis,

I don’t quite understand your comment “If you use Strings as the keys you won’t 
get affinity collocation set up properly”.

I have an object with a plain String key, named switchId, and another object 
with a composite key, of which one field is switchId (of type String). I am 
using switchId as the affinity key, and it seems to work fine without 
distributed queries. Was this a coincident and I should re-test?

My understanding is that caches with no explicit affinity key use the key to 
distribute objects. For the plain-key cache this would be switchId.

So, if I have two caches, both with switchId as the key, I should get automatic 
collocation (no setup required). Correct?

And for the third cache, with the switchId field in the composite key as the 
affinity key, I should also get collocation (with the other two caches). 
Correct?

What am I missing?

BTW, I think there was a typo in the response (first sentence), and you meant 
to say “set up properly and _non_-distributed joins will return an incomplete 
result.”.

Roger

PS: My configuration (all XML):

        <property name="cacheKeyConfiguration">
            <list>
                <!-- ports cache -->
                <bean class="org.apache.ignite.cache.CacheKeyConfiguration">
                    <property name="typeName" 
value="com.abc.poc.inppoc.model.PortKey"/>
                    <property name="affinityKeyFieldName" value="switchId"/>
                </bean>
                <!-- switches cache uses default affinity key (id) -->
            </list>
        </property>



                <!-- ports cache -->
                <bean 
class="org.apache.ignite.configuration.CacheConfiguration">
                    <property name="name" value="PortCache"/>
                    <property name="cacheMode" value="PARTITIONED"/>
                    <property name="backups" value="1"/>

                    <!-- use affinity (on Switch ID) -->
                    <property name="affinity">
                        <bean 
class="org.apache.ignite.cache.affinity.rendezvous.RendezvousAffinityFunction">
                            <property name="partitions" value="1024"/>
                        </bean>
                    </property>

                    <!-- Port fields/columns -->
                    <property name="queryEntities">
                        <list>
                            <bean class="org.apache.ignite.cache.QueryEntity">
                                <property name="keyType" 
value="com.abc.poc.inppoc.model.PortKey"/>
                                <property name="valueType" 
value="com.abc.poc.inppoc.model.Port"/>
                                <property name="fields">
                                    …


                <!-- switches cache -->
                <bean 
class="org.apache.ignite.configuration.CacheConfiguration">
                    <property name="name" value="SwitchCache"/>
                    <property name="cacheMode" value="PARTITIONED"/>
                    <property name="backups" value="1"/>

                    <!-- use affinity (on Switch ID, ie primary key) -->
                    <property name="affinity">
                        <bean 
class="org.apache.ignite.cache.affinity.rendezvous.RendezvousAffinityFunction">
                            <property name="partitions" value="1024"/>
                        </bean>
                    </property>

                    <!-- Switch fields/columns -->
                    <property name="queryEntities">
                        <list>
                            <bean class="org.apache.ignite.cache.QueryEntity">
                                <property name="keyType" 
value="java.util.UUID"/>
                                <property name="valueType" 
value="com.brocade.poc.inppoc.model.Switch"/>
                                <property name="fields">
                                    …

PortKey.java:

public class PortKey implements Serializable {

    private UUID    id;        // port-id; PK
    private UUID    switchId;  // affinity key; not really part of PK

    …



From: Denis Magda [mailto:dma...@apache.org]
Sent: Friday, September 01, 2017 3:07 PM
To: user@ignite.apache.org
Subject: Re: SQLQuery with simple Join return no results

If you use Strings as the keys you won’t get affinity collocation set up 
properly and distributed joins will return an incomplete result. One of the 
keys have to comprise a “parent” class key that will be an affinity key. Look 
at the example here:
https://apacheignite.readme.io/docs/affinity-collocation#section-collocate-data-with-data<https://urldefense.proofpoint.com/v2/url?u=https-3A__apacheignite.readme.io_docs_affinity-2Dcollocation-23section-2Dcollocate-2Ddata-2Dwith-2Ddata&d=DwMFaQ&c=IL_XqQWOjubgfqINi2jTzg&r=1esZO0r0bYS90lcsaLA6N4AFxuNo6lzauhETGwdJQoQ&m=KTlx1Z0HYXpMSZKonCG-plYFaijMpCt0OUS3tJgKbWQ&s=-YW5SFh-ko4-_EBwKFMKH5BOJU7pGcQMqaCRZrwZz24&e=>

As for the NON collocated joins suggested by Roger (qry.setDistributedJoins( 
true)), I would use them only if it’s impossible to set up the collocation 
between 2 entities. That’s not your case from what I see. NON collocated joins 
are slower than collocated ones.

—
Denis


On Sep 1, 2017, at 2:53 PM, Roger Fischer (CW) 
<rfis...@brocade.com<mailto:rfis...@brocade.com>> wrote:

Hi Matt,

are the objects to join collocated, ie. do they have the same affinity key? If 
yes, it should work (it worked for me).

If no, you need to enable distributed joins for the query. See the middle line.

           SqlFieldsQuery qry = new SqlFieldsQuery( stmt);
           qry.setDistributedJoins( true);
           queryCursor = aCache.query( qry);

Roger

-----Original Message-----
From: matt [mailto:goodie...@gmail.com]
Sent: Friday, September 01, 2017 1:52 PM
To: user@ignite.apache.org<mailto:user@ignite.apache.org>
Subject: SQLQuery with simple Join return no results

I have 2 caches defined, both with String keys, and classes that make use of 
the Ignite annotations for indexes and affinity. I've got 3 different nodes 
running, and the code I'm using to populate the cache w/test data works, and I 
can see each node is updated with its share of the data. My index types are set 
on the caches as well.

If I do a ScanQuery, I can see that all of the fields and IDs are correct, 
Ignite returns them all. But when doing a SqlQuery, I get nothing back.
Ignite is not complaining about the query, it's just returning an empty cursor.

If I remove the Join, results are returned.

So I'm wondering if this is related to the way I've set up my affinity mapping. 
It's basically setup like the code below... and the query looks like this:

"from B, A WHERE B.id = A.bID"

Any ideas on what I'm doing wrong here?

class A implements Serializable {
 @QuerySqlField(index = true)
 String id;

 @QuerySqlField(index = true)
 String bId;

 @AffinityKeyMapped
 @QuerySqlField(index = true)
 String group;
}

class B implements Serializable {
 @QuerySqlField(index = true)
 String id;

 @AffinityKeyMapped
 @QuerySqlField(index = true)
 String group;
}



--
Sent from: 
https://urldefense.proofpoint.com/v2/url?u=http-3A__apache-2Dignite-2Dusers.70518.x6.nabble.com_&d=DwICAg&c=IL_XqQWOjubgfqINi2jTzg&r=1esZO0r0bYS90lcsaLA6N4AFxuNo6lzauhETGwdJQoQ&m=g9B7o3WZd7LuW39MRFWdU5Gim1S3DRPZGcugI0n9Wno&s=ehBogkdblpG8U9N0taPx5Jdf8G9uDMnZiRR-f34rEe0&e=

Reply via email to