[orientdb] OrientDB community edition on AWS Cloud

2016-01-22 Thread MuthuKumar Kubendran
Hi All, Here is my fundamental question. Please reply if anyone has successfully enabled distributed architecture on AWS cloud by having ELB and Auto scaling with community edition Would it be possible to use distributed mode and clusters on AWS enabling the services elastic load

[orientdb] Re: Query with a combination of LET and Subselect is very slow

2016-01-22 Thread Hung Tran
Hi Alessandro, OMG, it gives a better result. 2.5 sec instead. Look at its explain, I don't know why it is faster. With that query style, how could I write a similar query like SELECT Statistics, $active[0].count, $total[0].count FROM ContactList LET $active = (SELECT count(*) FROM Contact

[orientdb] Re: How to speed up a query with non-indexable field in WHERE

2016-01-22 Thread Luca Son
Hi Hung Tran, I understand, unfortunately it is not currently possible to create indexes across multiple classes. Could you try with this SB-TREE index ? CREATE INDEX EntityInfo.State ON EntityInfo (State) NOTUNIQUE instead of your HASH_INDEX CREATE INDEX EntityInfo.State ON EntityInfo

[orientdb] Re: Query with a combination of LET and Subselect is very slow

2016-01-22 Thread Luca Son
Hi Hung Tran, yes, if it were possible it would help a lot. Thanks Luca -- --- You received this message because you are subscribed to the Google Groups "OrientDB" group. To unsubscribe from this group and stop receiving emails from it, send an email to

[orientdb] Re: How to speed up a query with non-indexable field in WHERE

2016-01-22 Thread Hung Tran
Hi Luca, I tried, the result is the same. I used HASH_INDEX because I don't need ORDER BY on that, so HASH_INDEX should give the best performance for WHERE. My Best, Hung Tran On Friday, January 22, 2016 at 4:30:16 PM UTC+7, Luca Son wrote: > > Hi Hung Tran, > > I understand, unfortunately it

[orientdb] Re: Query with a combination of LET and Subselect is very slow

2016-01-22 Thread Luca Son
Hi, thanks for the DB. To compare the results, could you tell me what's your best execution time about the above query ? Thanks Luca -- --- You received this message because you are subscribed to the Google Groups "OrientDB" group. To unsubscribe from this group and stop receiving emails

[orientdb] Re: How to speed up a query with non-indexable field in WHERE

2016-01-22 Thread Hung Tran
Hi Luca, The system does not allow me to do that, CREATE INDEX Contact.EntityInfo_State ON Contact (EntityInfo.State) NOTUNIQUE will generate an error com.orientechnologies.orient.core.index.OIndexException: Index with name : 'Contact.EntityInfo_State' cannot be created on class : 'Contact'

[orientdb] Re: Query with a combination of LET and Subselect is very slow

2016-01-22 Thread Hung Tran
Hi Luca, Here you are CREATE CLASS Contact CREATE PROPERTY Contact.Status INTEGER ALTER PROPERTY Contact.Status MANDATORY true ALTER PROPERTY Contact.Status NOTNULL true CREATE INDEX Contact.Status ON Contact (Status) NOTUNIQUE_HASH_INDEX CREATE PROPERTY Contact.EmailAddress STRING

Re: [orientdb] struggling to Migrate orientdb 1.7.2 to 2.2.0

2016-01-22 Thread Luca Garulli
Hi Jayaraj, We are working to make things easier and more coherent in OrientDB. Some "non-regular" behavior that worked in the past couldn't work anymore with recent version. For example you shouldn't never use select out_friends.name from user But rather the out() function. If we'll change

[orientdb] Result set of vertices with an empty collection of edges, vertices

2016-01-22 Thread Athanassios Hatzis
Hi, this might be a simple question but I am confused, please help. I am using OrientDB 2.1.9 and I am experimenting with VehicleHistoryGraph database. >From Studio, Browse mode, set limit to 9 records only. Now I am entering this simple query select out() from Person The result set I

[orientdb] Re: Result set of vertices with an empty collection of edges, vertices

2016-01-22 Thread Athanassios Hatzis
Yes it does help, thank you On Friday, 22 January 2016 15:47:54 UTC+2, Athanassios Hatzis wrote: > > Hi, > this might be a simple question but I am confused, please help. > > I am using OrientDB 2.1.9 and I am experimenting with VehicleHistoryGraph > database. > From Studio, Browse mode, set

Re: [orientdb] Result set of vertices with an empty collection of edges, vertices

2016-01-22 Thread Luigi Dell'Aquila
Hi Athanassios, This is the expected behavior, but you can obtain what your are looking for just adding a filter to the query: select out() from Person where out().size() > 0 Consider that out() is a normal projection, you could have more than one, eg. select name as personName, out() from

[orientdb] Re: Query with a combination of LET and Subselect is very slow

2016-01-22 Thread alessandrorota04
Hi, can you try with this query ? SELECT OwnedBy[EntityInfo.State=0][Account=#20:0].Statistics, count(*) FROM Contact where Status = 0 group by OwnedBy Best regards, Alessandro -- --- You received this message because you are subscribed to the Google Groups "OrientDB" group. To unsubscribe

[orientdb] OrientDB Minimum hardware requirements

2016-01-22 Thread Denis Wilson Souza Rosa
Hi Guys! What are the OrientDB minimum hardware requirements? I know it is a tough question, specially because OrientDB does not require anything special other than java 1.6 but this is a question that I always got from clients. Thanks -- --- You received this message because you are

[orientdb] Re: OrientDB Minimum hardware requirements

2016-01-22 Thread user . work108
Hi, it really depends on what the db is used for; basically, according to official guide, the only requirement is a system running java 1.7 or higher Il giorno venerdì 22 gennaio 2016 18:38:03 UTC+1, Denis Wilson Souza Rosa ha

[orientdb] Re: Query with a combination of LET and Subselect is very slow

2016-01-22 Thread Hung Tran
Hi Luca, 3s is the best result on my VM with that query (I don't know why JVM process is staying at 27% until now, that may impact on the result) My Best, Hung Tran On Friday, January 22, 2016 at 5:22:13 PM UTC+7, Luca Son wrote: > > Hi, > > thanks for the DB. > To compare the results, could