zero323 commented on a change in pull request #27278:
[SPARK-30569][SQL][PYSPARK][SPARKR] Add percentile_approx DSL functions.
URL: https://github.com/apache/spark/pull/27278#discussion_r380145375
##########
File path: sql/core/src/main/scala/org/apache/spark/sql/functions.scala
##########
@@ -652,6 +652,50 @@ object functions {
*/
def min(columnName: String): Column = min(Column(columnName))
+ /**
+ * Aggregate function: Returns and array of the approximate percentile values
+ * of numeric column col at the given percentages.
+ *
+ * Each value of the percentage array must be between 0.0 and 1.0.
+ *
+ * The accuracy parameter is a positive numeric literal
+ * which controls approximation accuracy at the cost of memory.
+ * Higher value of accuracy yields better accuracy, 1.0/accuracy
+ * is the relative error of the approximation.
+ *
+ * @group agg_funcs
+ * @since 3.1.0
+ */
+ def percentile_approx(e: Column, percentage: Array[Double], accuracy: Long):
Column = {
Review comment:
Huh... Usability aside, could you clarify one thing for me. Percentiles have
to be provided, as a literal. Even assuming that prevailing attitude is
"exception is good enough" when non literal is passed
> AnalysisException: cannot resolve ... due to data type mismatch: The
accuracy or percentage provided must be a constant literal;
However, arguments that require SQL literals:
- `ignoreNulls` in `first`
- `ignoreNulls` in `last`
- ` rsd`in `approx_count_distinct`
not to count different args in strict window functions are exclusively
passed as external literals not `Columns`\*. So what is the rationale behind
diverging here? What makes this case different?
It doesn't seem to be particularly verbose, two signatures are on the low
side for overloaded functions with multiple arguments and I don't see any
reason to anticipate significant maintenance effort (docs or otherwise).
I've seen some echos of discussions about growing size of
`o.a.s.sql.functions`, but these were hardly convincing, at least in the
diluted form they reached me.
---------------------------
* In fact there are quite a few functions, where only external literal are
passed, despite planner accepting any expression.
----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
For queries about this service, please contact Infrastructure at:
[email protected]
With regards,
Apache Git Services
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
