search

Azure / azure-cosmosdb-java

Java Async SDK for SQL API of Azure Cosmos DB
MIT License
54 stars 61 forks source link

Duplicate results when DISTINCT querying #367

Open MiDoWD opened 3 years ago

MiDoWD commented 3 years ago

Describe the bug DISTINCT query via the 2.6.11 Java async SDK is returning duplicate results.

To Reproduce Via the 2.6.11 Java async SDK I am querying documents with query which looks like

SELECT DISTINCT o.nest1.nest2.field as id FROM Objects o

with

    FeedOptions queryOptions = new FeedOptions();
    queryOptions.setMaxItemCount(10);
    queryOptions.setEnableCrossPartitionQuery(true);

The database has ~750 entries with 40 distinct values of this key. However after iterating all results from the pages I get, there are 69 of them, containing some duplicate ids. When I run the same query via the Portal UI, I get the expected 40 results. When I change the setMaxItemCount to 1000, I get the expected 40 results.

Not sure whether it's relevant, but the database uses this field o.nest1.nest2.field as partition key.

Expected behavior DISTINCT query should not return repeated entries.

Actual behavior Repeated entries are seen (69 entries for 40 distinct values).

Environment summary SDK Version: 2.6.11 (first seen on 2.6.3, update did not help) Java JDK version: 1.8.0_15 (same seen on JDK 11) OS Version: seen both on MacOS and Linux.

Additional context