pymilvus Search Results

GuideLines

Module Contents

class milvus.client.abstract.TopKQueryResult(raw_source, **kwargs)[source]

Bases: object

TopK query results, shown as 2-D array

This Class unpack response from server, store ids and distances separately.

property distance_array

Distance array, it's a 2-D array

property id_array

Id array, it's a 2-D array.

property raw

getter. return the raw result response

property shape

getter. return result shape, format as (row, column).

class milvus.client.abstract.RowQueryResult(_id_list, _dis_list)[source]

Bases: object

class milvus.client.abstract.QueryResult(_id, _distance)[source]

Bases: object

How to deal with search results

We provide two ways to deal with results. For examples, when you want to get result which position is (i, j), id and distance can be visited by:

  • id = results[i][j].id, distance = results[i][j].distance

  • id = results.id_array[i][j], distance = results.distance_array[i][j]

And you can ergodic whole results by:

>>> for row in results:
>>>     for item in row:
>>>         print("id={}, distance={}".format(item.id, item.distance))

or

>>> for id_list, dis_list in zip(results.id_array, results.distance_array):
>>>     for id, dis in zip(id_list, dis_list):
>>>        print("id={}, distance={}".format(id, dis))

We test the two way and record process time. Note that we comment print(...) sentence when testing. In situation of searching with nq=10000 and topk=1000, the first way cost about 8.88s and second way only 0.37s. If you want your program to run faster, the second way is recommended.