Filtering is not searching.
Filtering means that you reduce a data set by certain criteria and you deal with all the records that match the criteria.
Searching means that you keep the originally loaded data and scan the data for certain criteria and then jump from one hit to the next.
When I filter data, I prepare the data for further treatment. I do not need any highlight as all the records would be highlighted somehow as they all match the criteria. Otherwise the record would not be included in the list.
Filtering is very powerful for huge amounts of records. Searching is absolutely tedious for huge amounts of records as you may jump through hundreds or even thousands of hits.
The data that is compiled with filtering may be thousands of records apart. So it would take hours of scrolling to find the next highlighted entry.
In addition I see technical and logical problems with the filter hightlights. THis may be straightforward for simple filter criteria but tricky for more complex criteria or criteria that refer to fields that are not displayed or criteria that test for non-existing data (e.g. %bpm% MISSING).
There has been a discussion in 2010 on a similar topic:
/t/10820/1
The response back then has been just as reluctant as this time. So apparently there is no urgent need for this.