add filter and subset to documentation #2900
Conversation
| - when using `.` (`getpropery`) notation (`df.A`), | ||
| - when a single row is selected using an integer (`df[1, [:A, :B]]`) | ||
| - when `view` or `@view` is used (e.g. `@view df[1:3, :A]`). | ||
|
|
||
| More details on copies, views, and references can be found | ||
| in the [`getindex` and `view`](@ref) section. | ||
|
|
||
| An alternative approach to row subsetting in a data frame is to use | ||
| [`filter`](@ref), [`filter!`](@ref), [`subset`](@ref), or [`subset!`](@ref) |
There was a problem hiding this comment.
Are you sure we should advertise filter now that we have subset? The version taking a function is slow due to type instability, and the other one is similar to subset.
There was a problem hiding this comment.
filter is a bit faster. It is also easier to use in simple cases as you do not have to do broadcasting or using ByRow wrapper. But we can discuss this of course.
There was a problem hiding this comment.
OK. But then at least maybe don't show the pair syntax with filter? Better only show it for functions which interpret it the same way (subset, transform...).
There was a problem hiding this comment.
You mean that at this stage of the tutorial it would be too confusing to discuss it? In this case I would remove filter and only describe subset and just add a comment that there is also filter and that the reader can check it in the documentation. OK?
Co-authored-by: Milan Bouchet-Valat <[email protected]>
Co-authored-by: Milan Bouchet-Valat <[email protected]>
|
Thank you! PS. Julia is easier than English 😄. |
This is a long standing missing part of the docs.