[
https://issues.apache.org/jira/browse/SPARK-58044?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Rakesh Raushan updated SPARK-58044:
-----------------------------------
Description:
This sub-task is part of the umbrella SPARK-57550 (extend support for the TIME
data type).
AVG() expects NumericType,YearMonthIntervalType and DayTimeIntervalType as
input. It throws error for Time datatype as it assumes Time datatype is not
quantifiable.
Time is internally stored as a long value only which can be quantified and
average can be computed for time datatype as well.
Other aggregate functions like min(), max(), percentile(), approx_percentile(),
etc are already supported.
Currently if a user wants to calculate average on time column, user has to cast
it to numeric type, calculate average and then cast back to time type.
Supporting TIME datatype columns as native types in AVG() would allow users to
write cleaner and error free code.
was:
AVG() expects NumericType,YearMonthIntervalType and DayTimeIntervalType as
input. It throws error for Time datatype as it assumes Time datatype is not
quantifiable.
Time is internally stored as a long value only which can be quantified and
average can be computed for time datatype as well.
Other aggregate functions like min(), max(), percentile(), approx_percentile(),
etc are already supported.
Currently if a user wants to calculate average on time column, user has to cast
it to numeric type, calculate average and then cast back to time type.
Supporting TIME datatype columns as native types in AVG() would allow users to
write cleaner and error free code.
> Support TIME datatype in AVG()
> ------------------------------
>
> Key: SPARK-58044
> URL: https://issues.apache.org/jira/browse/SPARK-58044
> Project: Spark
> Issue Type: Sub-task
> Components: SQL
> Affects Versions: 4.3.0
> Reporter: Rakesh Raushan
> Priority: Major
>
> This sub-task is part of the umbrella SPARK-57550 (extend support for the
> TIME data type).
> AVG() expects NumericType,YearMonthIntervalType and DayTimeIntervalType as
> input. It throws error for Time datatype as it assumes Time datatype is not
> quantifiable.
> Time is internally stored as a long value only which can be quantified and
> average can be computed for time datatype as well.
> Other aggregate functions like min(), max(), percentile(),
> approx_percentile(), etc are already supported.
>
> Currently if a user wants to calculate average on time column, user has to
> cast it to numeric type, calculate average and then cast back to time type.
> Supporting TIME datatype columns as native types in AVG() would allow users
> to write cleaner and error free code.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
