[FEA] Add null_count_if_available
or lazy_null_count
to column
and column_view
#12256
Labels
feature request
New feature or request
improvement
Improvement / enhancement to an existing function
Performance
Performance related issue
The classes
column
andcolumn_view
havenull_count
functions that return the number of nulls in the given column. The column holds an internal null count value which is defaulted toUNKNOW_NULL_COUNT
.During various transformations, we need to copy that null count value around but may not really care about what that value is. However, if the function
null_count()
is called but the column's internal number of nulls has not been computed, a kernel will always be launched to compute the actual null count. That is a waste of time because we can just returnUNKNOW_NULL_COUNT
without compute anything.I suggest adding a new function that allows returning the column's internal null count directly without computation. That function can be named
null_count_if_availabe
orlazy_null_count
or anything else that is relevant.The text was updated successfully, but these errors were encountered: