[
https://issues.apache.org/jira/browse/IGNITE-20478?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Pavel Pereslegin updated IGNITE-20478:
--------------------------------------
Description:
Currently, when scanning an index, we set a special value called
"UNSPECIFIED_VALUE_PLACEHOLDER" to row. Which means that any value matches the
bound (more details in IGNITE-16443).
To be able to complete the transition to using a binary tuple, we need to
rework this approach and try to avoid storing non-conforming schema values in
row.
Currently, this placeholder sets to row when the search bound is open (that is,
when the RexNode is null in the list when creating a scalar).
{{ExpressionFactoryImpl#expandBounds}} needs to be reworked and there should be
no open bounds (see {{ExpressionFactoryImpl#compile}} all nodes elements must
not be null).
After reworking {{expandBounds}} the {{searchRow}} that comes to
{{RowConverter#toBinaryTuplePrefix}} should already contain a prefix only.
In {{ExpressionFactoryImpl#comparator}} this placeholder does not seem to be
used and this code can be removed.
was:
Currently, when scanning an index, we set a special value called
"UNSPECIFIED_VALUE_PLACEHOLDER" to row. Which means that any value matches the
bound (more details in IGNITE-16443).
To be able to complete the transition to using a binary tuple, we need to
rework this approach and try to avoid storing non-conforming schema values in
row.
Currently, this placeholder is set to row when the search bound is open (that
is, when the RexNode is null in the list when creating a scalar).
{{ExpressionFactoryImpl#expandBounds}} needs to be reworked and there should be
no open bounds (see {{ExpressionFactoryImpl#compile}} all nodes elements must
not be null).
After reworking {{expandBounds}} the {{searchRow}} that comes to
{{RowConverter#toBinaryTuplePrefix}} should already contain a prefix only.
In {{ExpressionFactoryImpl#comparator}} this placeholder does not seem to be
used and this code can be removed.
> Sql. Rework use of UNSPECIFIED_VALUE_PLACEHOLDER in row.
> --------------------------------------------------------
>
> Key: IGNITE-20478
> URL: https://issues.apache.org/jira/browse/IGNITE-20478
> Project: Ignite
> Issue Type: Improvement
> Components: sql
> Reporter: Pavel Pereslegin
> Priority: Major
> Labels: ignite-3
>
> Currently, when scanning an index, we set a special value called
> "UNSPECIFIED_VALUE_PLACEHOLDER" to row. Which means that any value matches
> the bound (more details in IGNITE-16443).
> To be able to complete the transition to using a binary tuple, we need to
> rework this approach and try to avoid storing non-conforming schema values in
> row.
> Currently, this placeholder sets to row when the search bound is open (that
> is, when the RexNode is null in the list when creating a scalar).
> {{ExpressionFactoryImpl#expandBounds}} needs to be reworked and there should
> be no open bounds (see {{ExpressionFactoryImpl#compile}} all nodes elements
> must not be null).
> After reworking {{expandBounds}} the {{searchRow}} that comes to
> {{RowConverter#toBinaryTuplePrefix}} should already contain a prefix only.
> In {{ExpressionFactoryImpl#comparator}} this placeholder does not seem to be
> used and this code can be removed.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
