[Bug]: partial_update not working

[IvanHahan]

Is there an existing issue for this?

• I have searched the existing issues

Describe the bug

when doing upsert, partial_update parameter is recognized, but not working.

my schema (simplified) is

• id
• keywords
• text
• meta

I'm doing upsert wtih document {"id": "some_existing_id", "meta": "some_meta"}, and pass partial_update=True.

I get

pymilvus.exceptions.MilvusException: <MilvusException: (code=1100, message=fieldSchema(keywords) has no corresponding fieldData pass in: invalid parameter)>

milvus expects to add fields that are missing (that I dont want to be updated)

Expected Behavior

I expect partial update to work. so in the example above I can update meta without an error

Steps/Code To Reproduce behavior

Environment details

- Hardware/Softward conditions (OS, CPU, GPU, Memory):
- Method of installation (Docker, or from source):
- Milvus version (v0.3.1, or v0.4.0):
- Milvus configuration (Settings you made in `server_config.yaml`):

Anything else?

No response

[yhmo]

Seems no problem, the following example is a partial_update demo, it works well.
Which version of your milvus server and pymilvus client?

import random

from pymilvus import (
    MilvusClient, DataType, 
)

client = MilvusClient(uri="http://localhost:19530")
print(client.get_server_version())

COLLECTION_NAME = "AAA"

dim = 4
def gen_embedding():
    return [random.random() for k in range(dim)]


if __name__ == '__main__':
    client.drop_collection(collection_name=COLLECTION_NAME)
    schema = MilvusClient.create_schema(enable_dynamic_field=True)
    schema.add_field(field_name="id", datatype=DataType.INT64, is_primary=True, auto_id=False)
    schema.add_field(field_name="vector", datatype=DataType.FLOAT_VECTOR, dim=dim)
    schema.add_field(field_name="keywords", datatype=DataType.VARCHAR, max_length=100)
    schema.add_field(field_name="text", datatype=DataType.VARCHAR, max_length=100)
    schema.add_field(field_name="meta", datatype=DataType.VARCHAR, max_length=100)

    index_params = client.prepare_index_params()
    index_params.add_index(
        field_name="vector",
        index_type="FLAT",
        metric_type="COSINE",
    )
    client.create_collection(collection_name=COLLECTION_NAME,
                             schema=schema,
                             index_params=index_params,
                             )

    data = [
        {"id": 0, "vector": gen_embedding(), "meta": "A", "keywords": "000", "text": "text_0"},
        {"id": 1, "vector": gen_embedding(), "meta": "B", "keywords": "111", "text": "text_1"},
        {"id": 2, "vector": gen_embedding(), "meta": "C", "keywords": "222", "text": "text_2"},
    ]
    res = client.insert(collection_name=COLLECTION_NAME, data=data)

    print("Before upsert")
    res = client.query(collection_name=COLLECTION_NAME,
                       filter="id >= 0",
                       output_fields=["*"],
                       consistency_level="Strong")
    for r in res:
        print(r)

    client.upsert(collection_name=COLLECTION_NAME,
                  data={"id": 1, "meta": "updated"},
                  partial_update=True,
                  )
    print("After upsert")
    res = client.query(collection_name=COLLECTION_NAME,
                       filter="id >= 0",
                       output_fields=["*"],
                       consistency_level="Strong")
    for r in res:
        print(r)

Output:

Before upsert
{'text': 'text_0', 'meta': 'A', 'id': 0, 'vector': [0.2591438889503479, 0.8062371015548706, 0.9139207005500793, 0.6515963077545166], 'keywords': '000'}
{'text': 'text_1', 'meta': 'B', 'id': 1, 'vector': [0.5132219791412354, 0.566847026348114, 0.7745850086212158, 0.8649949431419373], 'keywords': '111'}
{'text': 'text_2', 'meta': 'C', 'id': 2, 'vector': [0.08315187692642212, 0.7454366087913513, 0.7430552840232849, 0.6156575083732605], 'keywords': '222'}
After upsert
{'keywords': '000', 'text': 'text_0', 'meta': 'A', 'id': 0, 'vector': [0.2591438889503479, 0.8062371015548706, 0.9139207005500793, 0.6515963077545166]}
{'keywords': '111', 'text': 'text_1', 'meta': 'updated', 'id': 1, 'vector': [0.5132219791412354, 0.566847026348114, 0.7745850086212158, 0.8649949431419373]}
{'keywords': '222', 'text': 'text_2', 'meta': 'C', 'id': 2, 'vector': [0.08315187692642212, 0.7454366087913513, 0.7430552840232849, 0.6156575083732605]}

[yhmo]

Note that partial upsert requires the milvus server version >= v2.6.2
And the pymilvus v2.6.2 release note mentioned it fixed some bugs about partial update: https://github.com/milvus-io/pymilvus/releases/tag/v2.6.2