[Bug]: auto index not working on text match search, results in Assertion error

[pycui]

Is there an existing issue for this?

• I have searched the existing issues

Environment

- Milvus version: 2.5.0-beta
- Deployment mode(standalone or cluster): cluster
- MQ type(rocksmq, pulsar or kafka):    pulsar
- SDK version(e.g. pymilvus v2.0.0rc2): 2.5.0
- OS(Ubuntu or CentOS): Ubuntu
- CPU/Memory: 64/250T
- GPU: none
- Others:

Current Behavior

When trying to search text field using text match, returns a queryNode error

MilvusException: <MilvusException: (code=65535, message=fail to Query on QueryNode 1467: worker(1467) query failed: Operator::GetOutput failed for [Operator:PhyFilterBitsNode, plan node id: 180] : Assert "iter != text_indexes_.end()"  => failed to get text index, text index not found at /workspace/source/internal/core/src/segcore/SegmentInterface.cpp:399
)>

schema

fields = [
    FieldSchema(name="id", dtype=DataType.INT64, is_primary=True, auto_id=True),
    FieldSchema(name="embedding", dtype=DataType.FLOAT_VECTOR, dim=1024)
]
collection_schema = CollectionSchema(fields, description="A collection with text and embedding vector")

collection_schema.add_field(
    field_name='text', 
    datatype=DataType.VARCHAR, 
    max_length=1000, 
    enable_analyzer=True, # Whether to enable text analysis for this field
    enable_match=True # Whether to enable text match
)

relevant index

index_params.add_index(
    field_name="text",
    index_type="", 
    index_name="text_index"
)

search code

filter = "TEXT_MATCH(text, 'sample text 1')"

result = client.query(
    collection_name="text_embedding_collection",
    filter=filter, 
    output_fields=["id", "text"]
)

using embedding + text match hybrid also gets the same error

filter = "TEXT_MATCH(text, 'sample text 1')"
query_vector = np.random.random(1024).tolist()

result = client.search(
    collection_name="text_embedding_collection", 
    anns_field="embedding", 
    data=[query_vector], 
    filter=filter,
    search_params={"params": {"nprobe": 10}},
    limit=10, # Max. number of results to return
    output_fields=["id", "text"] # Fields to return
)

Expected Behavior

Returns retrieved rows

Steps To Reproduce

This can be reproduced by creating a new collection like above.

Milvus Log

No response

Anything else?

No response

[pycui]

I think the issue is we have to explicitly define index type, even though https://milvus.io/docs/scalar_index.md#Scalar-Index says we have Auto Indexing

[yanliang567]

/assign @zhengbuqian
/unassign

[SpadeA-Tang]

Do you insert any values or what are your insert queries? I used the relevant version in standalone cluster and have not reproduced it. @pycui

[SpadeA-Tang]

collection_schema.add_field(
field_name='text',
datatype=DataType.VARCHAR,
max_length=1000,
enable_analyzer=True, # Whether to enable text analysis for this field
enable_match=False # Whether to enable text match
)

When I set enable_match be false, I can reproduce the panic. So would you mind ensuring that you set this be True?
@pycui

[xiaofan-luan]

I think the issue is we have to explicitly define index type, even though https://milvus.io/docs/scalar_index.md#Scalar-Index says we have Auto Indexing

match seems to has nothing to do with index type? @pycui can you verify on 2.5.2 we released today?