Timeline / interest for not yet implemented foverlaps() functionality

[sritchie73]

There are a number of arguments in the foverlaps() function that are not yet implemented:

• [x] type = "equal" - #3416
• [ ] maxgap
• [ ] minoverlap

Along with a few requested features:

• [ ] #979 non integer support for minoverlap
• [ ] #1355 foverlaps() should work without setkey() beforehand
• [ ] #1975 return length of overlap
• [ ] #2645 foverlaps() could work with factor levels

There will be a lesson on the foverlaps() function in the DataCamp course on joins, so I'm wanting to gauge whether its worth trying to implement these features before the course is released. Are there obvious technical barriers that have prevented these features from being implemented? Or has there been a lack of interest from the community and developers, e.g. given the availability of GenomicRanges?

[msummersgill]

Related: #3007 foverlaps error message could be more informative

[arunsrinivasan]

@sritchie73 thanks for this. I wrote foverlaps in about 3-4 hrs (for EARL conference) and have not since looked at the features that've been left out .. Will take a look.

[Atrebas]

Some potentially helpful other references: bedops, bedtools, tidygenomics.

[Al-Murphy]

Hi, just wondering if maxgap, minoverlap is likely to be implemented soon? I think it would be a brilliant addition to avoid the conversion to GRanges object.

[MichaelChirico]

Hey @Al-Murphy that's definitely in "contributions welcome" status... happy to support however I can if you wanted to take up the gauntlet :smiley:

cc @arunsrinivasan who would be most familiar with that part of the code base & maybe better able to offer "mentor" support (not sure availability)

[jangorecki]

For those looking for not yet implemented functionality in foverlaps, please go to https://github.com/Rdatatable/data.table/issues/528#issuecomment-1880007663 and provide example code