[Answer]-Can you do this with a database?

An alternative way to do something like this would be to use the GROUP_CONCAT function across the columns that differ, and group on the column(s) that repeat(s). This will create a delimited string column in the output that you will need to parse in PHP; however, it will reduce your row count to only the number of unique combinations of the non-grouped columns. For example:

select name, playlist, GROUP_CONCAT(file, '|', quality, '|', size, '|', type SEPARATOR '/') AS delimited_pairs
from video
group by name, playlist;

In the above example, the resulting delimited_pairs column will delimit each unique column value per name/playlist (the grouping columns) using the pipe character, and each of these unique combinations will be delimited by forward slash.