ing by timestamp range,
>>>>> since
>>>>> the translation mechanism already exists.
>>>>>
>>>>> * Incremental scan using timestamp range is part of the SQL Standard
>>>>> (e.g. "VERSIONS BETWEEN SYSTEM TIME ... AND ...") and is supported by some
>>>>> existing engines. See [3] for SQL Server support.
>>>>>
>>>>> * Conceptually, it is possible to implement the same query semantics
>>>>> at the SQL level using existing APIs and operators such as "TIMESTAMP AS
>>>>> OF" and "EXCEPT" (e.g., by selecting: (T as of timestamp1) EXCEPT (T as of
>>>>> timestamp2)). It sounds that incremental scan is an optimization to push
>>>>> the differencing operation to the data source as opposed to letting the
>>>>> engine deal with it, so it is better to implement the SQL shorthand, and
>>>>> its corresponding data source optimization.
>>>>>
>>>>> Therefore, I would like to ask if we can proceed with supporting an
>>>>> API along the lines of "VERSIONS BETWEEN SYSTEM TIME ... AND ...", and
>>>>> consider the discussion in [1] as a caveat that folks using this API (or
>>>>> existing timestamp APIs) have to keep in mind, until the implementation
>>>>> guarantees linearity at some point in the future.
>>>>>
>>>>> Thanks,
>>>>> Walaa.
>>>>>
>>>>> [1] https://www.mail-archive.com/dev@iceberg.apache.org/msg01504.html
>>>>> [2]
>>>>> https://github.com/apache/iceberg/blob/1a11038baaca9f4cf03674654a6d9a85f39a1ce1/core/src/main/java/org/apache/iceberg/util/SnapshotUtil.java#L265
>>>>> [3]
>>>>> https://docs.microsoft.com/en-us/sql/relational-databases/tables/temporal-tables?view=sql-server-ver15
>>>>>
>>>>
>>>>
>>>> --
>>>> Ryan Blue
>>>> Tabular
>>>>
>>>
>
> --
> Ryan Blue
> Tabular
>
--
---
Best Regards
Peidian Li
+1， we haven the same needs, hope the solution to this problem. Thanks.
Can you also please invite me to join the slack. Thanks.
Congratulations!
