[orientdb] Dijkstra function may not work with 'weight' field on OrientDB-1.7-rc2

I want to gain shortest path with weight(ex. length of link). I'm trying dijkstra() function on OrientDB-1.7-rc2, but dijkstra() may not work. This function parameter is Dijkstra(shource_RID, target_RID, 'weight', 'direction'). Dijkstra() function work well on OrintDB-1.3. *Insert java code :* O

[orientdb] Re: Database does not shut down with lucene using Java API

But there is a difference to a deadlock. All code after graph.shutdown() is still going on. graph.shutdown() System.out.println("done"); The system.out is still executed after i called graph.shutdown(). Its strange. Am Mittwoch, 21. Mai 2014 06:26:01 UTC+2 schrieb Montrazul: > > Hallo, > > got

[orientdb] Database does not shut down with lucene using Java API

Hallo, got a really nasty problem. When i create a lucene index and i call graph.shutdown() the database never shuts down. The program never terminates and the database never frees the locks on the files. I always have to crash the program/database. public static void main(String[] args) { Stri

[orientdb] Re: SQL : projects and *

On a related note: The following is not true in the wiki: https://github.com/orientechnologies/orientdb/wiki/SQL-Query#projections "By using the wildcard * the fields retrieved will be the same but the RID and version will be lost:" In 1.7-SNAPSHOT: The following two queries yield the same result

[orientdb] SQL : projects and *

Can OrientDB support the following: select in_knows.name, * from V Currently (1.7-SNAPSHOT) this is interpreted as: select * from V This makes it hard to efficiently get all of the properties from each record without knowing their names (i.e. each record may have different properties). -- -

[orientdb] NullPointerException when doing OSQLAsynchQuery

Hello, I downloaded the orientdb-develop branch this morning. I'm issuing the following via remote mode to an empty database: OSQLAsynchQuery SELECT flatten(indexes) from cluster:0 The three results are returned (dictionary, OUser, ORole), and then a Null Pointer exception is thrown on com.ori

[orientdb] Re: Javascript JSON Marshalling

JSON.parse fixed this. On Tuesday, 20 May 2014 23:40:36 UTC+1, Thomas Kennedy wrote: > > I have a server side with javascript function one parameter called Data. > > The value of Data is {"test":5}. This is passed in as a string. > > in the javascript function I want to access the value of test

Re: [orientdb] Multiple fields in Lucene FULLTEXT index and embedded maps

Hi Enrico, Would it be possible for me to use the Lucene integration build my own flavor/version of the "document being indexed"? I'm not sure if this request will make sense to you but I would like more control over what gets put into the indexed-document to allow for search based on the Mult

[orientdb] Javascript JSON Marshalling

I have a server side with javascript function one parameter called Data. The value of Data is {"test":5}. This is passed in as a string. in the javascript function I want to access the value of test. return Data.test is not working either is var temp = JSON.stringify(Data); return temp.test

Re: [orientdb] Multiple fields in Lucene FULLTEXT index and embedded maps

Hi Enrico, Thank you. I have not been indexing Embedded maps so far but I would like to do that now with the fulltext option. Regards, - Stefán On Tuesday, 20 May 2014 18:02:42 UTC, Enrico Risa wrote: > > Hi Stefan > > for multiple fields you could create it like a normal composite index > >

Re: [orientdb] Configure Lucene to use KeywordAnalyzer using Java API

Seems like doesn't found the constructor with Version. I will check it out I've used it with the english analyzer. i will try the Keyword 2014-05-20 20:05 GMT+02:00 Montrazul : > Great! > But thats not working for the KeywordAnalyzer isnt it? > > ODocument metadata = new ODocument(); > metadata.

Re: [orientdb] Configure Lucene to use KeywordAnalyzer using Java API

Great! But thats not working for the KeywordAnalyzer is it? ODocument metadata = new ODocument(); metadata.field("analyzer", "org.apache.lucene.analysis.core.KeywordAnalyzer"); vertexType.createIndex(indexName, "FULLTEXT", null, metadata, "LUCENE", new String[] {fieldName}); java.lang.NoSuchM

Re: [orientdb] Configure Lucene to use KeywordAnalyzer using Java API

Great! But thats not working for the KeywordAnalyzer isnt it? ODocument metadata = new ODocument(); metadata.field("analyzer", "org.apache.lucene.analysis.core.KeywordAnalyzer"); vertexType.createIndex(indexName, "FULLTEXT", null, metadata, "LUCENE", new String[] {fieldName}); java.lang.NoSuc

Re: [orientdb] Multiple fields in Lucene FULLTEXT index and embedded maps

Hi Stefan for multiple fields you could create it like a normal composite index create index V.name_song_type on V (name,song_type) FULLTEXT ENGINE LUCENE and then you can query it with select from V where [name,song_type] LUCENE "diddley" For fields with sub-properties how do you index them n

Re: [orientdb] OrientDB use as a triple store

So I made some changes to allow loadRDF to take in a custom ValueFactory so that I can have shorter strings, for example: "http://example.com/rel"; -> "ex:rel" but now I still get similar errors now when adding edges. It works if the label has a period, so now I use: "ex:.rel". *Why does Orien

[orientdb] Multiple fields in Lucene FULLTEXT index and embedded maps

Hi, Can someone please outline the best practices for creating a full text search index on multiple fields and fields with sub-properties. Regards, -Stefán -- --- You received this message because you are subscribed to the Google Groups "OrientDB" group. To unsubscribe from this group and

Re: [orientdb] Configure Lucene to use KeywordAnalyzer using Java API

Great! Am Dienstag, 20. Mai 2014 19:30:55 UTC+2 schrieb Enrico Risa: > > Yes you can use it from the java api > > see the code here: > You can use the ODocument metadata in the createIndex api > > http://pastebin.com/Xrk2qPyT > > > 2014-05-20 19:19 GMT+02:00 Montrazul > >: > >> Hi, >> >> here: Fu

Re: [orientdb] Configure Lucene to use KeywordAnalyzer using Java API

Yes you can use it from the java api see the code here: You can use the ODocument metadata in the createIndex api http://pastebin.com/Xrk2qPyT 2014-05-20 19:19 GMT+02:00 Montrazul : > Hi, > > here: Full Text > Index >

[orientdb] Configure Lucene to use KeywordAnalyzer using Java API

Hi, here: Full Text Index is written how to create an index in orientdb using lucene. There is also an example how to configure the analyzer to be used by lucene: create index City.name on City (name) FULLTEXT ENGINE LU

Re: [orientdb] cluster support

Hi Dmitry, the cluster support will remain forever. What OrientDB doesn't allow anymore is to create a document of class X against a cluster Y if the cluster Y wasn't configured to host documents of class X. All the rest will work. Against clusters we designed the sharding architecture, so multipl

Re: [orientdb] Re: Record ID string JSON serialization

I see. So what would be the best approach to copy the ID of the vertex just created into another property named UUID ? On Tuesday, May 20, 2014 1:29:39 AM UTC-7, Artem Orobets wrote: > > Hi Amit, > > An explicit call toString() may brake your graph consistency. > > You are doing the changes in tr

[orientdb] cluster support

http://stackoverflow.com/questions/22606111/orientdb-problems-with-clusters-concept Hi! Do I understand correctly that support multiple clusters for the class will be canceled? It is best not to use this functionality when developing applications? Thanks. -- --- You received this message be

Re: [orientdb] [Orient 1.7.rc2] Bug in fulltext index

Yes i found that for console. but i didnt find out how to use it for java api yet Am Dienstag, 20. Mai 2014 15:05:54 UTC+2 schrieb Enrico Risa: > > Yes we have a METADATA field when creating the index and you can pass a > different analyzer. > > As i wrote here, i used the EnglishAnalyzer > > ht

Re: [orientdb] [Orient 1.7.rc2] Bug in fulltext index

Yes we have a METADATA field when creating the index and you can pass a different analyzer. As i wrote here, i used the EnglishAnalyzer https://github.com/orientechnologies/orientdb-lucene/wiki/Full-Text-Index 2014-05-20 14:20 GMT+02:00 Montrazul : > Will there maybe be an option to configure

Re: [orientdb] [Orient 1.7.rc2] Bug in fulltext index

Will there maybe be an option to configure if lucene should use the standardanalyzer or the keyanalyzer? Am Dienstag, 20. Mai 2014 14:14:29 UTC+2 schrieb Artem Orobets: > > Oh sorry, I meant it is not implemented right now. But I've added comment > to the bug to implement it. > > Best regards, >

Re: [orientdb] [Orient 1.7.rc2] Bug in fulltext index

Oh sorry, I meant it is not implemented right now. But I've added comment to the bug to implement it. Best regards, Artem Orobets * Orient Technologiesthe Company behind OrientDB* 2014-05-20 11:59 GMT+03:00 Riccardo Tasso : > I mean: can I pass to FULLTEXT index some configuration at creation

Re: [orientdb] Vertices with multiple Clusters

This is the database shape (as you wrote, I typed 'info in the console'): Current database: ClusterTest (url=remote:localhost/ClusterTest) Cluster configuration: { } DATABASE PROPERTIES: ++ NAME

Re: [orientdb] how to manage 1-N relation.

Sure: public class benchOrient { public static void main(String[] args) throws UnsupportedEncodingException, NoSuchAlgorithmException, FileNotFoundException{ int n=1000; double alpha=0.7; int maxtab=7; ODatabaseDocumentTx db = new ODatabaseDocumentTx("remote:/localhost/OrientBench").open("admin

Re: [orientdb] [Orient 1.7.rc2] Bug in fulltext index

I mean: can I pass to FULLTEXT index some configuration at creation time? I can't see any documentation about it. Thanks, Riccardo 2014-05-20 10:37 GMT+02:00 Artem Orobets : > It seems it is reasonable for all apostrophe usages, so why not > > Best regards, > Artem Orobets > > *Orient Technol

Re: [orientdb] [Orient 1.7.rc2] Bug in fulltext index

It seems it is reasonable for all apostrophe usages, so why not Best regards, Artem Orobets *Orient Technologies the Company behind OrientDB* 2014-05-20 11:11 GMT+03:00 Riccardo Tasso : > Is it possible to configure the FULLTEXT index to create two keys for > apostrophed keywords? > For exampl

Re: [orientdb] Re: Oriendb Lucene

Can i configure lucene to use KeywordAnalyzer instead of StandardAnalyzer? If you for example search for the word "log" with StandardAnalyzer and you have also "log1", "log2", ... in your index lucene will also return these results. With the KeywordAnalyzer he would just return the results for "

Re: [orientdb] Re: Record ID string JSON serialization

Hi Amit, An explicit call toString() may brake your graph consistency. You are doing the changes in transaction so ids of the elements are temporary and will be updated during transaction commit. Explicit call toString will prevent such update. This also mean that you need to commit transaction

Re: [orientdb] [Orient 1.7.rc2] Bug in fulltext index

Is it possible to configure the FULLTEXT index to create two keys for apostrophed keywords? For example "Juliet's" will be indexed both as "juliet", both as "juliets". Or at least indexed only as "juliet". Thanks, Riccardo 2014-05-20 10:08 GMT+02:00 Riccardo Tasso : > Yes it does, > here's t

Re: [orientdb] [Orient 1.7.rc2] Bug in fulltext index

Yes it does, here's the bug report: https://github.com/orientechnologies/orientdb/issues/2358 2014-05-20 10:01 GMT+02:00 Artem Orobets : > Hi Riccardo, > > Looks like a bug. Probably the reason is an apostrophe, does it works with > 'chamber' key? > Could you create a bug > report

Re: [orientdb] [Orient 1.7.rc2] Bug in fulltext index

Hi Riccardo, Looks like a bug. Probably the reason is an apostrophe, does it works with 'chamber' key? Could you create a bug report ? Best regards, Artem Orobets * Orient Technologiesthe Company behind OrientDB* 2014-05-20 10:53 GMT+03

[orientdb] [Orient 1.7.rc2] Bug in fulltext index

Hi, I'll try to explain briefly what seems a bug to me: I have a hierarchy of classes, but I want a fulltext index (case insensitive) over all of them for the same field "label", here I create: alter property V.label COLLATE ci; create index V-label-fulltext ON V (label) FULLTEXT; Then I try the