"Compiling token bank fails with error" - for CREATE LOADING JOB

GSQL Loading
1

I am trying to create a graph and was able to successfully define the node and edge schemas and define the graph, as can be seen in the output of the ‘ls’ command

GSQL > ls
---- Graph gsb_tigergraph
Vertex Types: 
  - VERTEX gsbNode(node_id UINT primary key, locx FLOAT, locy FLOAT, region STRING, node_name STRING) WITH STATS="NONE"
Edge Types: 
  - UNDIRECTED EDGE gsbEdge(FROM gsbNode, TO gsbNode, source_id UINT, target_id UINT, weight FLOAT, flag BOOL)

Graphs: 
  - Graph gsb_tigergraph(gsbNode:v, gsbEdge:e)
Jobs: 
Queries:

However the next step of creating a loading job fails

GSQL > create loading job load_node_data for graph gsb_tigergraph {define filename fileVar; load fileVar to vertex gsbNode values ($0, $1, $2, $3, $4) using separator=",", header="true", eol="\n";}
Compiling token bank fails with error
Failed to create loading jobs: [load_node_data].

I am using TigerGraph community version 4.2.0 on Ubuntu 24.04. Yes, I know TigerGraph is not yet supported on Ubuntu 24.04, but I was able to bypass the OS version check in the installation scripts.

When can we expect TigerGraph to be supported on Ubuntu 24.04?

2

Hi @ameyap2 ,

Not sure if you still have the issue…

Can you try to delete the graph and then use the PRIMARY_ID instead of the PRIMARY KEY?

E.g. initialize the graph with the following command:

CREATE GRAPH gsb_tigergraph()

BEGIN
USE GRAPH gsb_tigergraph()

CREATE SCHEMA_CHANGE JOB init_gsb_graph FOR GRAPH gsb_tigergraph {
    ADD VERTEX gsbNode(PRIMARY_ID node_id UINT, locx FLOAT, locy FLOAT, region STRING, node_name STRING) WITH PRIMARY_ID_AS_ATTRIBUTE="true";
    ADD UNDIRECTED EDGE gsbEdge(FROM gsbNode, TO gsbNode, source_id UINT, target_id UINT, weight FLOAT, flag BOOL)
}
END

And then you can re-create the loading job again:

BEGIN
CREATE LOADING JOB load_node_data for graph gsb_tigergraph {
    define filename fileVar;
    load fileVar to vertex gsbNode values ($0, $1, $2, $3, $4) using separator=",", header="true", eol="\n";
}
END

Best,
Supawish Limprasert (Jim)
Solution Engineer, TigerGraph

3

Changing the PRIMARY KEY to PRIMARY_ID did not help either

GSQL > CREATE VERTEX gsbNode (PRIMARY_ID node_id UINT, locx FLOAT, locy FLOAT, region STRING, node_name STRING) WITH STATS="none", WITH PRIMARY_ID_AS_ATTRIBUTE="true"
Successfully created vertex types: [gsbNode].

GSQL > CREATE UNDIRECTED EDGE gsbEdge (FROM gsbNode, TO gsbNode, source_id UINT, target_id UINT, weight FLOAT, flag BOOL)
Successfully created edge types: [gsbEdge].

GSQL > create graph gsb_tigergraph (gsbNode, gsbEdge)
Compiling token bank fails with error
Waiting for GPE and GSE online...
The graph gsb_tigergraph is created.

GSQL > CREATE LOADING JOB load_node_data FOR GRAPH gsb_tigergraph { DEFINE FILENAME file_gsbNode=<file_path>; LOAD file_gsbNode TO VERTEX gsbNode VALUES($"node_id", $"locx", $"locy", $"region", $"node_name") USING SEPARATOR=",", HEADER="true", EOL="\n";}
Compiling token bank fails with error
Failed to create loading jobs: [load_node_data].
4

Hi @ameyap2 ,

Sorry if that didn’t work out.

IF you are using TG Community version via docker, you might have to make sure that you configured the disk space correctly. Or, you might want to search the TigerGraph log files for the errors above: Log Files :: TigerGraph DB

If you are a current client of TigerGraph, please reach out to the TigerGraph support team so we can take a further look into the issue.

Best,
Supawish Limprasert (Jim)
Solution Engineer, TigerGraph