Skip to content

Document aggregation pushdown in Oracle connector #11805

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
hashhar merged 1 commit into from
Apr 7, 2022
Merged

Document aggregation pushdown in Oracle connector #11805

Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
23 changes: 23 additions & 0 deletions docs/src/main/sphinx/connector/oracle.rst
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -398,6 +398,29 @@ The connector supports pushdown for a number of operations:
* :ref:`limit-pushdown`
* :ref:`topn-pushdown`

In addition, the connector supports :ref:`aggregation-pushdown` for the
following functions:

* :func:`avg()`
* :func:`count()`, also ``count(distinct x)``
* :func:`max()`
* :func:`min()`
* :func:`sum()`

Pushdown is only supported for ``DOUBLE`` type columns with the
following functions:
Comment on lines +410 to +411
Copy link
Copy Markdown
Member

@hashhar hashhar Apr 7, 2022

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Some such type limitations also apply to the functions above as well but they can vary based on type mappings and what the actual data-type is in remote system.

Longer term I would prefer to move docs to a direction where we clearly set the expectation that operations will be pushed down where possible without trying to overspecify those conditions since it's complex, will never get 100% accurate due to the amount of possible combinations and will change without much visibility.

Copy link
Copy Markdown
Member Author

@mosabua mosabua Apr 7, 2022

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

That sounds good .. the problem is that users test this stuff and then ask specific questions why something is and is not pushed down .. ideally the docs have the answer..

Copy link
Copy Markdown
Member

@hashhar hashhar Apr 7, 2022 •
edited
Loading

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

The answer is if it doesn't work then it's unsupported. We obviously need a better way to communicate that for sure.

For a taste of complexity to know whether sum(x) will be pushed down:

  • data type of x as Trino sees it
  • the underlying data type in remote database
  • any implicit casts that get added by engine (may vary based on query shape)
  • the data type of the grouping columns used
  • whether x is an expression or a plain column reference
  • whether there are other operations that cannot be pushed down (this is unbounded list of things)

As an example of practical real world case where people don't really care about "when" things get pushed down is that sum() since the beginning has only been pushdown-able if the column is a bigint or a decimal but no one ever complained or asked about this. People generally care about performance - not exactly how that performance is achieved.


* :func:`stddev()` and :func:`stddev_samp()`
* :func:`stddev_pop()`
* :func:`var_pop()`
* :func:`variance()` and :func:`var_samp()`

Pushdown is only supported for ``REAL`` or ``DOUBLE`` type column
with the following functions:

* :func:`covar_samp()`
* :func:`covar_pop()`

.. _oracle-predicate-pushdown:

Predicate pushdown support
Expand Down