Dropping a collection permanently deletes the collection and everything stored in it (vectors, metadata fields, indexes, partitions). This action is irreversible.

Be careful: If you need the data later, export/backup it before you drop the collection.

Drop a collection

Use drop_collection() to remove a collection by name.

from dodil import Client
from dodil.vbase import VBaseConfig
 
# Authorize using your Service Account
c = Client(
    service_account_id="...",
    service_account_secret="...",
)
 
# Connect to your VBase database
vbase = c.vbase.connect(
    VBaseConfig(
        host="vbase-db-<id>.infra.dodil.cloud",
        port=443,
        scheme="https",
        db_name="db_<id>",
    )
)
 
collection_name = "my_collection"
 
# (Optional) guard: check it exists first
if not vbase.has_collection(collection_name):
    raise RuntimeError(f"Collection '{collection_name}' does not exist")
 
# Drop the collection (irreversible)
vbase.drop_collection(collection_name)
 
print("Dropped:", collection_name)

Verify it was removed

After dropping, you can list collections (or check existence) to confirm.

print(vbase.list_collections())
print("exists?", vbase.has_collection("my_collection"))

Notes and best practices

• Irreversible: Dropping deletes data and indexes. If you need to keep the data, export or copy it to another collection first.
• Active workloads: If your application is actively querying/searching the collection, you may see transient errors until clients refresh. Consider pausing traffic before dropping.
• Loaded collections: If the collection is currently loaded in memory, dropping it still removes the collection, but it’s a good operational habit to release it first in busy environments:

vbase.release_collection("my_collection")
vbase.drop_collection("my_collection")

• Permissions: Your Service Account must have permission to manage collections for the target database.