Running Loading Jobs

Once a job is created, you execute it using the RUN LOADING JOB command.

1. Execution Command

Basic syntax to run a job:

bashterminal
Copy
RUN LOADING JOB my_job

Overriding File Paths

You can change the data source at runtime using the USING clause:

bashterminal
Copy
RUN LOADING JOB my_job USING f="/new/path/data.csv"

2. Runtime Options

Enhance your loading process with these flags:

• -dryrun: Processes data but does not write to the graph. Useful for validation.
• -noprint: Suppresses the progress output; only shows the Job ID and log path.
• -n [i,j]: Limits loading to a specific range of lines (e.g., -n 1,100).
• -max_error me: Aborts the job if the number of failed lines exceeds me.

3. Maintenance Commands

Clearing Data

• CLEAR GRAPH STORE: Flushes all data from the database but preserves the schema.
  • Use -HARD to bypass confirmation: CLEAR GRAPH STORE -HARD.
• DROP ALL: Wipes both data and schema definitions.

Deleting Specific Data

Use DELETE statements inside a loading job to remove specific vertices or edges based on input files:

gsqlterminal
Copy
CREATE LOADING JOB purge_job FOR GRAPH MyGraph {
  DEFINE FILENAME f;
  DELETE VERTEX Person (PRIMARY_ID $0) FROM f;
}

4. REST API Integration

For high-performance or automated environments, run jobs via the REST++ endpoint: POST /ddl/{graph_name}

This bypasses the GSQL shell overhead and is ideal for production pipelines.

5. Troubleshooting

• Check Status: Use gadmin status to ensure GPE and GSE services are online.
• Logs: Every job generates a log file (path provided at the start of the run). Check these for specific "Rejected Line" errors.