[
https://issues.apache.org/jira/browse/SPARK-56976?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Richard Durkee updated SPARK-56976:
-----------------------------------
Description:
Currently Spark's IF function requires exactly three arguments:
{code:java}
IF(condition, trueValue, falseValue){code}
The two-argument form, IF(condition, value), returns value when true and NULL
when false. This is supported by MySQL, Trino/Presto, and BigQuery but rejected
by Spark with an arity error.
Although there are existing ways to express this in Spark, such as
IF(condition, value, NULL) or CASE WHEN condition THEN value END, the
two-argument form is more concise and matches user expectations when migrating
from other SQL engines. This is especially important for automated migration.
was:
Currently Spark's IF function requires exactly three arguments:
{code:java}
IF(condition, trueValue, falseValue){code}
The two-argument form that many languages offer, IF(condition, value), returns
value when true and NULL when false. This is supported by MySQL, Trino/Presto,
and BigQuery but rejected by Spark with an arity error.
Although there are many existing ways to express this in Spark, such as
IF(condition, value, NULL) or CASE WHEN condition THEN value END, the
two-argument form is more concise and matches user expectations when migrating
from other SQL engines. This is especially important for automated migration.
> Support Two-Argument IF(condition, value) Returning NULL
> --------------------------------------------------------
>
> Key: SPARK-56976
> URL: https://issues.apache.org/jira/browse/SPARK-56976
> Project: Spark
> Issue Type: Improvement
> Components: SQL
> Affects Versions: 5.0.0
> Reporter: Richard Durkee
> Priority: Minor
> Labels: pull-request-available
>
> Currently Spark's IF function requires exactly three arguments:
> {code:java}
> IF(condition, trueValue, falseValue){code}
> The two-argument form, IF(condition, value), returns value when true and NULL
> when false. This is supported by MySQL, Trino/Presto, and BigQuery but
> rejected by Spark with an arity error.
> Although there are existing ways to express this in Spark, such as
> IF(condition, value, NULL) or CASE WHEN condition THEN value END, the
> two-argument form is more concise and matches user expectations when
> migrating from other SQL engines. This is especially important for automated
> migration.
>
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
