Method for S4 DataFrame for left_join

[lambdamoses]

This way we can use colData(sce) in left join into another SCE object without having to convert to S3 tibble.

[stemangiola]

Do you have in mind (?)

single_cell_experiment |> *_join(DataFrame)

It sounds good!

[lambdamoses]

Yes, that's what I mean. So we can do something like sce |> left_join(colData(sce2)).

[stemangiola]

Sure,

we would need all join flavours (full, right, left, anti), as tidySingleCellExperiment has API for all of them. But it should be pretty straightforward.

[lambdamoses]

Do you want me to do a pull request? I think it should be pretty easy to implement, though I won't be able to do it soon because I'm busy with many other things recently.

[stemangiola]

Do you want me to do a pull request? I think it should be pretty easy to implement, though I won't be able to do it soon because I'm busy with many other things recently.

That would be great.

[lambdamoses]

OK, super last minute for Bioc 3.18. Question: It seems that the code in those different flavors of *_join is all the same except for the dplyr function called. Do you want me to refactor using a function factory, so if you want to change the *_join functions, then you only need to change 1 place instead of 4 places?

[stemangiola]

OK, super last minute for Bioc 3.18. Question: It seems that the code in those different flavors of *_join is all the same except for the dplyr function called. Do you want me to refactor using a function factory, so if you want to change the *_join functions, then you only need to change 1 place instead of 4 places?

Would not https://github.com/lambdamoses/tidySingleCellExperiment/blob/bcd3f5529948e9cff2f317bbf001558db9986bae/R/dplyr_methods.R#L329

take all arguments of join*?

[lambdamoses]

Yes, it takes all arguments of *_join. The outer function controls the small parts that differ among the 4 join flavors, while the inner function implements the large part that is the same for all flavors of join. The outer function .join_factory returns the inner function, which takes all arguments of join.

Since you want S4 DataFrame for y for tidySE, do you want the refactored version?

[stemangiola]

the thing that confuses me is that left_join.SingleCellExperiment should take all the argument of dplyr left_join. but it does not in your refactored version

[lambdamoses]

What do you mean? The original code says, dplyr::left_join(y, by=by, copy=copy, suffix=suffix, ...), and the arguments that show up here and ... are preserved in .join_factory's inner function. The other arguments are passed through ....