Skip to main content

​
Method

client.batches.delete(
    name: str,
    config: Optional[DeleteBatchJobConfig] = None
) -> DeleteResourceJob
Deletes a batch job from your project. This operation removes the job metadata but does not delete output files that have already been written.
​
name
string
required
The resource name or ID of the batch job to delete:
  • Vertex AI: projects/{project}/locations/{location}/batchPredictionJobs/{job_id} or just the {job_id} if project/location are set in the client
  • Gemini API: batches/{batch_id} or just the {batch_id}
​
config
DeleteBatchJobConfig
Optional configuration for the delete request

​
Response

​
name
string
The resource name of the delete operation
​
done
boolean
Whether the deletion is complete
​
error
object
Error information if the deletion failed

​
Usage

​
Delete a Batch Job

from google import genai

client = genai.Client(api_key='your-api-key')

# Delete a batch job
delete_op = client.batches.delete(name='123456789')

print(f"Delete operation: {delete_op.name}")
print(f"Complete: {delete_op.done}")

​
Delete with Error Handling

from google import genai

client = genai.Client(api_key='your-api-key')

job_name = '123456789'

try:
    # Check if job exists first
    batch_job = client.batches.get(name=job_name)
    print(f"Found job: {batch_job.name}")
    
    # Delete the job
    delete_op = client.batches.delete(name=job_name)
    
    if delete_op.done:
        print(f"Successfully deleted job: {job_name}")
    else:
        print(f"Delete operation in progress: {delete_op.name}")
        
except Exception as e:
    print(f"Error deleting job: {e}")

​
Delete Multiple Jobs

from google.genai import types

# Delete all failed jobs
deleted_count = 0
error_count = 0

for batch_job in client.batches.list():
    if batch_job.state == types.JobState.JOB_STATE_FAILED:
        try:
            client.batches.delete(name=batch_job.name)
            print(f"Deleted: {batch_job.name}")
            deleted_count += 1
        except Exception as e:
            print(f"Error deleting {batch_job.name}: {e}")
            error_count += 1

print(f"\nDeleted {deleted_count} jobs")
if error_count > 0:
    print(f"Failed to delete {error_count} jobs")

​
Delete Old Jobs

from datetime import datetime, timedelta
from google.genai import types

# Delete jobs older than 30 days
thirty_days_ago = datetime.now() - timedelta(days=30)

deleted_jobs = []
for batch_job in client.batches.list():
    job_time = datetime.fromisoformat(
        batch_job.create_time.replace('Z', '+00:00')
    )
    
    if job_time < thirty_days_ago:
        # Only delete completed or failed jobs
        if batch_job.state in [
            types.JobState.JOB_STATE_SUCCEEDED,
            types.JobState.JOB_STATE_FAILED,
            types.JobState.JOB_STATE_CANCELLED
        ]:
            try:
                client.batches.delete(name=batch_job.name)
                deleted_jobs.append(batch_job.name)
                print(f"Deleted old job: {batch_job.name}")
            except Exception as e:
                print(f"Could not delete {batch_job.name}: {e}")

print(f"\nDeleted {len(deleted_jobs)} old jobs")

​
Conditional Deletion

from google.genai import types

job_name = '123456789'

# Check job state before deleting
batch_job = client.batches.get(name=job_name)

if batch_job.state in [
    types.JobState.JOB_STATE_SUCCEEDED,
    types.JobState.JOB_STATE_FAILED,
    types.JobState.JOB_STATE_CANCELLED
]:
    # Safe to delete completed jobs
    delete_op = client.batches.delete(name=job_name)
    print(f"Deleted completed job: {job_name}")
else:
    print(f"Cannot delete job in state: {batch_job.state}")
    print("Cancel the job first or wait for it to complete")

​
Cleanup After Processing Results

import json
from google.cloud import storage
from google.genai import types

job_name = '123456789'
batch_job = client.batches.get(name=job_name)

if batch_job.state == types.JobState.JOB_STATE_SUCCEEDED:
    # Download results first
    gcs_uri = batch_job.dest.gcs_uri
    print(f"Downloading results from {gcs_uri}...")
    
    # Download logic here...
    results = []  # Your downloaded results
    
    # Save results locally
    with open('batch_results.json', 'w') as f:
        json.dump(results, f)
    
    print("Results saved locally")
    
    # Now safe to delete the job
    client.batches.delete(name=job_name)
    print(f"Deleted batch job: {job_name}")

​
Bulk Cleanup

from google.genai import types

# Delete all completed jobs
terminal_states = [
    types.JobState.JOB_STATE_SUCCEEDED,
    types.JobState.JOB_STATE_FAILED,
    types.JobState.JOB_STATE_CANCELLED
]

print("Finding completed jobs...")
completed_jobs = []

for batch_job in client.batches.list():
    if batch_job.state in terminal_states:
        completed_jobs.append(batch_job)

print(f"Found {len(completed_jobs)} completed jobs")

if completed_jobs:
    response = input(f"Delete all {len(completed_jobs)} jobs? (yes/no): ")
    
    if response.lower() == 'yes':
        deleted = 0
        for job in completed_jobs:
            try:
                client.batches.delete(name=job.name)
                deleted += 1
                print(f"Deleted {deleted}/{len(completed_jobs)}: {job.name}")
            except Exception as e:
                print(f"Error: {e}")
        
        print(f"\nSuccessfully deleted {deleted} jobs")
    else:
        print("Deletion cancelled")

​
Important Notes

Deleting a batch job is permanent and cannot be undone. Make sure to download any results before deletion.
  • Deletion removes job metadata but NOT output files in GCS or BigQuery
  • You can delete jobs in any state, including running jobs
  • Deleting a running job does NOT cancel it - use batches.cancel first
  • Output files must be deleted separately if needed
  • Some deletion operations may be asynchronous (check done field)

​
What Gets Deleted

Deleted:
  • Job metadata and configuration
  • Job history and logs
  • Reference to input/output locations
Not Deleted:
  • Output files in GCS buckets
  • Output tables in BigQuery
  • Input files
  • Uploaded files in Files API

​
Error Handling

The delete operation may fail in the following cases:
  • Job not found: The specified job name doesn’t exist
  • Permission denied: Insufficient permissions to delete the job
  • Network error: Connection issues with the API
try:
    client.batches.delete(name=job_name)
    print("Job deleted successfully")
except ValueError as e:
    print(f"Invalid job name: {e}")
except PermissionError as e:
    print(f"Permission denied: {e}")
except Exception as e:
    print(f"Unexpected error: {e}")

​
See Also

Build docs developers (and LLMs) love

Get started for freeTalk to us