return array data type in infer_dtype_bydata

[rgupta2508]

Now milvus supporting ARRAY data types so we can return DataType.ARRAY in case of if it is not VECTOR.

This will be helpful to create ARRAY data type field seamlessly in milvus

[sre-ci-robot]

Welcome @rgupta2508! It looks like this is your first PR to milvus-io/pymilvus 🎉

[rgupta2508]

@wangting0128 @longjiquan Please review , if this can be merged to return Milvus.ARRAY data type for infer_dtype_bydata function.

Or any thing we can do.

[rgupta2508]

/assign @czs007

[sre-ci-robot]

[APPROVALNOTIFIER] This PR is NOT APPROVED

This pull-request has been approved by: rgupta2508
To complete the pull request process, please ask for approval from czs007 after the PR has been reviewed.

The full list of commands accepted by this bot can be found here.

Details

Needs approval from an approver in each of these files:

• OWNERS

Approvers can indicate their approval by writing /approve in a comment
Approvers can cancel approval by writing /approve cancel in a comment

[rgupta2508]

This PR will resolve this issue #2165
idea is to take kwargs which accept datatype along with all the parameter, based on kwargs we can decide return type.
As per current implementation it is difficult to find which one is vector array or float array

[longjiquan]

Please help to review this. @xiaocai2333
/assign @xiaocai2333

[xiaocai2333]

It is recommended to change the default behavior to infer the vector, as pymilvus itself relies on the interface.

[longjiquan]

@rgupta2508 thanks for the contributions. I agree that we should infer data as array when the data is list-like and the infered type is bool or varchar, but it's also not enough to infer the schema since element type is required for array.

[longjiquan]

@rgupta2508 Take array type as considerations, I prefer that this function should infer a field schema instead of a data type (since data type don't contain other informations).

[longjiquan]

If so, I think we should do much more refactors.

[rgupta2508]

@rgupta2508 thanks for the contributions. I agree that we should infer data as array when the data is list-like and the infered type is bool or varchar, but it's also not enough to infer the schema since element type is required for array.

Yes correct , we cant return field schema with this, this we can use for data type only. directly we cant use it for schema. even max_capacity for array and max_length for varchar array is also mandatory to create schema, not sure if we can use this method for this, this will just return DataType only.

used should handle this if they will get datatyepe Array.

[rgupta2508]

@rgupta2508 Take array type as considerations, I prefer that this function should infer a field schema instead of a data type (since data type don't contain other informations).

Yes correct, but even max_capacity for array and max_length for varchar array is also mandatory to create schema, not sure if we can add some default values on this.

[rgupta2508]

Hi @longjiquan Any suggestion on this PR . if this can be merged with this idea.
Similar issue previously #1785