I’m trying to run CREATE QUERY on a graph.
However, this causes the node running the graph to reach massive CPU usage for a very long time, and then the creation fails.
However, on a new cluster with an almost empty schema, creation and installation did succeed, though after considerable time. (So it’s not a syntax error)
There is no error message but “Internal Error”.
This is the basic outline of the query:
CREATE QUERY myquery(/* some params here */) FOR GRAPH mygraph {
/* accumualtors and variable declerations*/
Start = {sourcesEntities};
WHILE Start.size() > 0 LIMIT max_hops DO
@@currentHop += 1;
startSingleHop = SELECT t FROM Start:s -(:e)- :t
WHERE (/* conditions */)
ACCUM
/* ACCUM block logic here */
; #end of select
startTwoHops = Start MINUS startSingleHop;
IF (some_condition) THEN
startTwoHops = SELECT t FROM startTwoHops:s -(:e)- :t -(:enext)- :n
WHERE (/* conditions */)
ACCUM
/* ACCUM block logic here */
HAVING /* some condition */
; # end of select
END;
Start = startSingleHop UNION startTwoHops;
END;
PRINT results;
}
Note that the query is using V2 Syntax and Pattern Matching.
How can I know why creation fails?
Using TigerGraph 3.5.x