[SPARK-36920][SQL] Support ANSI intervals by ABS()
#34169
Closed
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
|
Kubernetes integration test starting |
|
Kubernetes integration test status failure |
MaxGekk
changed the title
ABS()ABS()
|
@beliefer @AngersZhuuuu @Peng-Lei @cloud-fan Could you review this PR, please. |
|
Test build #143814 has finished for PR 34169 at commit
|
HyukjinKwon
approved these changes
Oct 5, 2021
|
Merged to master. |
cloud-fan
reviewed
Oct 5, 2021
| @@ -160,11 +162,15 @@ case class Abs(child: Expression, failOnError: Boolean = SQLConf.get.ansiEnabled | |||
|
|
|||
| def this(child: Expression) = this(child, SQLConf.get.ansiEnabled) | |||
|
|
|||
| override def inputTypes: Seq[AbstractDataType] = Seq(NumericType) | |||
| override def inputTypes: Seq[AbstractDataType] = Seq(TypeCollection.NumericAndInterval) | |||
There was a problem hiding this comment.
We shouldn't use TypeCollection.NumericAndInterval, as it includes the legacy interval. We will hit runtime exception if we use legacy interval as input, but analysis exception is preferred.
There was a problem hiding this comment.
@cloud-fan Thanks for the comment. Here is the fix #34183
MaxGekk
added a commit
that referenced
this pull request
Oct 5, 2021
…ANSI intervals ### What changes were proposed in this pull request? Change allowed input types of `Abs()` from: ``` NumericType + CalendarIntervalType + YearMonthIntervalType + DayTimeIntervalType ``` to ``` NumericType + YearMonthIntervalType + DayTimeIntervalType ``` ### Why are the changes needed? The changes make the error message more clear. Before changes: ```sql spark-sql> set spark.sql.legacy.interval.enabled=true; spark.sql.legacy.interval.enabled true spark-sql> select abs(interval -10 days -20 minutes); 21/10/05 09:11:30 ERROR SparkSQLDriver: Failed in [select abs(interval -10 days -20 minutes)] java.lang.ClassCastException: org.apache.spark.sql.types.CalendarIntervalType$ cannot be cast to org.apache.spark.sql.types.NumericType at org.apache.spark.sql.catalyst.util.TypeUtils$.getNumeric(TypeUtils.scala:77) at org.apache.spark.sql.catalyst.expressions.Abs.numeric$lzycompute(arithmetic.scala:172) at org.apache.spark.sql.catalyst.expressions.Abs.numeric(arithmetic.scala:169) ``` After: ```sql spark.sql.legacy.interval.enabled true spark-sql> select abs(interval -10 days -20 minutes); Error in query: cannot resolve 'abs(INTERVAL '-10 days -20 minutes')' due to data type mismatch: argument 1 requires (numeric or interval day to second or interval year to month) type, however, 'INTERVAL '-10 days -20 minutes'' is of interval type.; line 1 pos 7; 'Project [unresolvedalias(abs(-10 days -20 minutes, false), None)] +- OneRowRelation ``` ### Does this PR introduce _any_ user-facing change? No, because the original changes of #34169 haven't released yet. ### How was this patch tested? Manually checked in the command line, see examples above. Closes #34183 from MaxGekk/fix-abs-input-types. Authored-by: Max Gekk <max.gekk@gmail.com> Signed-off-by: Max Gekk <max.gekk@gmail.com>
a0x8o
added a commit
to a0x8o/spark
that referenced
this pull request
Oct 5, 2021
…ANSI intervals ### What changes were proposed in this pull request? Change allowed input types of `Abs()` from: ``` NumericType + CalendarIntervalType + YearMonthIntervalType + DayTimeIntervalType ``` to ``` NumericType + YearMonthIntervalType + DayTimeIntervalType ``` ### Why are the changes needed? The changes make the error message more clear. Before changes: ```sql spark-sql> set spark.sql.legacy.interval.enabled=true; spark.sql.legacy.interval.enabled true spark-sql> select abs(interval -10 days -20 minutes); 21/10/05 09:11:30 ERROR SparkSQLDriver: Failed in [select abs(interval -10 days -20 minutes)] java.lang.ClassCastException: org.apache.spark.sql.types.CalendarIntervalType$ cannot be cast to org.apache.spark.sql.types.NumericType at org.apache.spark.sql.catalyst.util.TypeUtils$.getNumeric(TypeUtils.scala:77) at org.apache.spark.sql.catalyst.expressions.Abs.numeric$lzycompute(arithmetic.scala:172) at org.apache.spark.sql.catalyst.expressions.Abs.numeric(arithmetic.scala:169) ``` After: ```sql spark.sql.legacy.interval.enabled true spark-sql> select abs(interval -10 days -20 minutes); Error in query: cannot resolve 'abs(INTERVAL '-10 days -20 minutes')' due to data type mismatch: argument 1 requires (numeric or interval day to second or interval year to month) type, however, 'INTERVAL '-10 days -20 minutes'' is of interval type.; line 1 pos 7; 'Project [unresolvedalias(abs(-10 days -20 minutes, false), None)] +- OneRowRelation ``` ### Does this PR introduce _any_ user-facing change? No, because the original changes of apache/spark#34169 haven't released yet. ### How was this patch tested? Manually checked in the command line, see examples above. Closes #34183 from MaxGekk/fix-abs-input-types. Authored-by: Max Gekk <max.gekk@gmail.com> Signed-off-by: Max Gekk <max.gekk@gmail.com>
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
What changes were proposed in this pull request?
In the PR, I propose to handle ANSI interval types by the
Absexpression, and theabs()function as a consequence of that:ABS()returns the same input value,Int.MinValuemonths for year-month interval andLong.MinValuemicroseconds for day-time interval),ABS()throws the arithmetic overflow exception.ABS()negate its input and returns a positive interval.For example:
Why are the changes needed?
To improve user experience with Spark SQL.
Does this PR introduce any user-facing change?
No, this PR just extends
ABS()by supporting new types.How was this patch tested?
By running new tests: