dijkstra throwing exceptions in some cases

[agarciaGIT]

OrientDB Version: 2.2.29

Java Version: 1.8.0_111

OS: Amazon Linux AMI 2016.09

Expected behavior

Running dijkstra will return the cheapest path between two vertices

Actual behavior

Exceptions are thrown in certain cases

Steps to reproduce

I have a graph where I added the Property 'weight' to E. I created other Sub classes from E and V. If I run shortestPath I get a response. If I run dijkstra on the same source and destination I get an exception.

orientdb {db=SpinOffIntelligence}> select shortestPath(#30:5,#39:3,'BOTH')

+----+---------------------+
|# |shortestPath |
+----+---------------------+
|0 |[#30:5,#13:780,#39:3]|
+----+---------------------+

1 item(s) found. Query executed in 0.002 sec(s).
orientdb {db=SpinOffIntelligence}> select dijkstra(#30:5,#39:3,'weight','BOTH')

Error: com.orientechnologies.orient.core.exception.OCommandExecutionException: Error on execution of command: sql.select dijkstra(#30:5,#39:3,'weight','BOTH')
DB name="SpinOffIntelligence"
DB name="SpinOffIntelligence"

Running dijkstra with other Vertices does work some time. I can't tell why.

orientdb {db=SpinOffIntelligence}> select dijkstra(#30:5,#39:5,'weight','BOTH')

+----+-------------------------+
|# |dijkstra |
+----+-------------------------+
|0 |[#30:5,#31:0,#16:6,#39:5]|
+----+-------------------------+

[agarciaGIT]

SpinOffIntelligence (2).gz

[luigidellaquila]

HI @agarciaGIT

you have a broken DB, some edges point to non-existing vertices.
In my instance I found this edge

+----+-------+--------+-----+--------+------+----------------+
|# |@Rid |@Class |out |in |weight|type |
+----+-------+--------+-----+--------+------+----------------+
|0 |#20:129|EPartner|#39:6|#19:1798|1 |Research Partner|
+----+-------+--------+-----+--------+------+----------------+

(I don't know if the RID matches, during the import it can change) that points to #19:1798 vertex, that doesn't exist.

I'm adding some checks to dijkstra() function to prevent exceptions, but the real problem here is not in the implementation

Thanks

Luigi

[agarciaGIT]

OK thanks. I will drop the DB and reload it and see if that works. AG

…

On Mon, Nov 6, 2017 at 10:01 AM, luigidellaquila ***@***.***> wrote: HI @agarciaGIT <https://github.com/agarciagit> you have a broken DB, some edges point to non-existing vertices. In my instance I found this edge +----+-------+--------+-----+--------+------+----------------+ |# ***@***.*** <https://github.com/rid> ***@***.*** <https://github.com/class> |out |in |weight|type | +----+-------+--------+-----+--------+------+----------------+ |0 |#20 <#20> :129|EPartner|#39 <#39>:6|#19 <#19>:1798|1 |Research Partner| +----+-------+--------+-----+--------+------+----------------+ (I don't know if the RID matches, during the import it can change) that points to #19 <#19>:1798 vertex, that doesn't exist. I'm adding some checks to dijkstra() function to prevent exceptions, but the real problem here is not in the implementation Thanks Luigi — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub <#7869 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/ACQXMaQy65MmmIyabLLJ3ZXYVukqAFkmks5szxE-gaJpZM4QSAkq> .