mongodb/mongo-python-driver
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

PYTHON-4666 Fix handling of large documents in client.bulk_write (#1798)

Browse Source
This commit is contained in:
Shruti Sridhar 2024-08-19 12:57:57 -07:00 committed by GitHub
parent f16206cb89
commit ad888797cf
No known key found for this signature in database
GPG Key ID: B5690EEEBB952194
3 changed files with 6 additions and 8 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

2
pymongo/message.py
View File

@ -1123,6 +1123,8 @@ def _client_batched_op_msg_impl(
new_message_size = total_ops_length + total_ns_length + op_length + ns_length
# We have enough data, return this batch.
if new_message_size > max_doc_sequences_bytes:
if idx == 0:
_raise_document_too_large(op_type, op_length, max_bson_size + _COMMAND_OVERHEAD)
break
# Add op and ns documents to this batch.

6
test/asynchronous/test_client_bulk_write.py
View File

@ -512,17 +512,15 @@ class TestClientBulkWriteCRUD(AsyncIntegrationTest):
# Document too large.
b_repeated = "b" * self.max_message_size_bytes
models = [InsertOne(namespace="db.coll", document={"a": b_repeated})]
with self.assertRaises(InvalidOperation) as context:
with self.assertRaises(DocumentTooLarge):
await client.bulk_write(models=models)
self.assertIn("cannot do an empty bulk write", context.exception._message)
# Namespace too large.
c_repeated = "c" * self.max_message_size_bytes
namespace = f"db.{c_repeated}"
models = [InsertOne(namespace=namespace, document={"a": "b"})]
with self.assertRaises(InvalidOperation) as context:
with self.assertRaises(DocumentTooLarge):
await client.bulk_write(models=models)
self.assertIn("cannot do an empty bulk write", context.exception._message)
@async_client_context.require_version_min(8, 0, 0, -24)
@unittest.skipUnless(_HAVE_PYMONGOCRYPT, "pymongocrypt is not installed")

6
test/test_client_bulk_write.py
View File

@ -512,17 +512,15 @@ class TestClientBulkWriteCRUD(IntegrationTest):
# Document too large.
b_repeated = "b" * self.max_message_size_bytes
models = [InsertOne(namespace="db.coll", document={"a": b_repeated})]
with self.assertRaises(InvalidOperation) as context:
with self.assertRaises(DocumentTooLarge):
client.bulk_write(models=models)
self.assertIn("cannot do an empty bulk write", context.exception._message)
# Namespace too large.
c_repeated = "c" * self.max_message_size_bytes
namespace = f"db.{c_repeated}"
models = [InsertOne(namespace=namespace, document={"a": "b"})]
with self.assertRaises(InvalidOperation) as context:
with self.assertRaises(DocumentTooLarge):
client.bulk_write(models=models)
self.assertIn("cannot do an empty bulk write", context.exception._message)
@client_context.require_version_min(8, 0, 0, -24)
@unittest.skipUnless(_HAVE_PYMONGOCRYPT, "pymongocrypt is not installed")