Implement boolean filter types in the Data Explorer #332
Conversation
| @@ -144,6 +144,14 @@ | |||
| nzchar(col) | |||
There was a problem hiding this comment.
Should I be expecting an update to filter_params in .ps.filter_rows()?
Even if no params, at least to show we are handling this case?
There was a problem hiding this comment.
A good question ... That mapping explicitly only contains the names of the operations that take parameters. Parameter free operations like "is null" are excluded.
There was a problem hiding this comment.
Ah I missed that bit thanks
| @@ -144,6 +144,14 @@ | |||
| nzchar(col) | |||
| } | |||
|
|
|||
| .ps.filter_col.is_true <- function(col, params) { | |||
| col == TRUE | |||
There was a problem hiding this comment.
I assume this is only applied on logical columns? i.e. could col ever not be a logical vector? (I would assume that since the frontend knows the column types, it should only present the user with applicable filter options)
Also, what do you want it to return in the NA case? NA == TRUE returns NA, so if the return value is allowed to be all 3 of true/false/na then that works I guess. I just immediately assumed it should be a vector of either TRUE or FALSE for what gets returned to the frontend.
If col is always logical and NAs are ok, a simplification for these would be to just return col and !col respectively.
If we want to remove NAs from the picture, col & !is.na(col) and !col & !is.na(col) would work.
There was a problem hiding this comment.
The output of this is ultimately fed to which, so NA is fine, but I agree removing them would be cleaner.
There was a problem hiding this comment.
(also, yes, the UI takes care of ensuring that these filter types only get applied to logical columns)
This change adds support for the
is_trueandis_falsefilter types for the R data explorer backend. See posit-dev/positron#2949 for details.