Dear R Developers, ### # Context: ###
When managing Search Path Conflicts (See: https://developer.r-project.org/Blog/public/2019/03/19/managing-search-path-conflicts/index.html), with: options(conflicts.policy = "strict") We get the following behaviour when loading a package (Eg: dplyr): library(dplyr) ## Error: Conflicts attaching package ‘dplyr’: ## ## The following objects are masked from ‘package:stats’: ## ## filter, lag ## ## The following objects are masked from ‘package:base’: ## ## intersect, setdiff, setequal, union So we would have to solve the conflict by writing: library(dplyr, mask.ok = c("filter", "lag", "intersect", "setdiff", "setequal", "union")) So my feature request proposals: ### # Feature Request 1: Interactive Session ### Would it be possible to raise an input prompt, which asks user for an action to be taken as regards conflicts? This would make the package loading process more dynamic when being loaded for first time in an interactive session (Eg: R Notebook). An example: The first time the package is loaded: options(conflicts.policy = "strict", conflicts.policy.ask = TRUE) library(dplyr) Executes iteratively the code, in order to ask the user for action (See toy example): opt <- readline(prompt="1: mask.ok; 2: exclude. Choose: ") if (opt == "1"){ txt <- sprintf( fmt = "conflictRules(pkg = '%s' , mask.ok = '%s')", "package.name", "variable.name" ) eval(parse(text = txt)) message(txt) } else if(opt == "2"){ txt <- sprintf( fmt = "conflictRules(pkg = '%s' , exclude = '%s')", "package.name", "variable.name" ) eval(parse(text = txt)) message(txt) } And afterwards, a message is printed with the selected "conflictRules()" Configuration for the dynamic setup. The user will only have to put the printed pre-configured "conflictRules()" setup into his code, and when re-executing, the input prompt asking for action will not be raised back again. Such behaviour is similar in spirit to how 'conflicted' R package works, which prints for example, for dplyr::filter : Error: [conflicted] `filter` found in 2 packages. Either pick the one you want with `::` * dplyr::filter * stats::filter Or declare a preference with `conflict_prefer()` * conflict_prefer("filter", "dplyr") * conflict_prefer("filter", "stats") Where the user will only have to copy "conflict_prefer("filter", "dplyr")" and paste it into his script. The difference, is that with: options(conflicts.policy = "strict", conflicts.policy.ask = TRUE), such message is printed at package load. I would put the required code within the library() function with the following conditional: ... if (length(conflicts)) { if(getOption("conflicts.policy.ask") == TRUE){ ... } } ... ### # Feature Request 2: Source Execution ### Another alternative, which could apply when executing the .R file from source, would be to suggest the user a default conflictRules() setup: options(conflicts.policy = "strict") library(dplyr) # Error: Conflicts attaching package ‘dplyr’: # # The following objects are masked from ‘package:stats’: # # filter, lag # # The following objects are masked from ‘package:base’: # # intersect, setdiff, setequal, union # # Declare preference with `conflictRules()` before loading: # * conflictRules("dplyr", mask.ok = list(stats = TRUE, base = TRUE)) # * conflictRules("dplyr", mask.ok = list(stats = c("filter", "lag"), base = c("intersect", "setdiff", "setequal", "union"))) # * conflictRules("dplyr", exclude = c("filter", "lag", "intersect", "setdiff", "setequal", "union")) In this case, the error message would have to be extended with sensible suggested defaults. Thanks, Juan ______________________________________________ R-devel@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-devel
