Subset of data based on dynamic column names
I have a dataset called DF. I want to filter DF based on column 1, so I:
DFb <- subset(DF, Column1 == "ABC")
Now I want Column1 to be dynamically defined based on user input, so the user can filter based on any column. Something like that:
Column_Name = 'Column3' ## User defines this variable
DFb <- subset(DF, as.name(Column_Name) == "ABC")
However, this doesn't work. Any suggestions?
source to share
Use [[
-subsetting:
DFb <- DF[DF[[Column_Name]] == "ABC",]
It's not as elegant as it is subset()
, but it works. subset()
uses "custom evaluation", which is great for interactive use, but makes it difficult when you want to make this second-order link.
The main thing is [[
; you can use subset(DF,DF[[Column_Name]]=="ABC")
instead, the results will be (almost) equivalent ( subset()
automatically discards values where criterion evaluates to NA
...)
You can do it in a package dplyr
, which gives a lot of flexibility in preventing out-of-the-box evaluation, but it's still a bit circular (maybe the best way to do this: I'm not very experienced with dplyr
).
library("dplyr") ## for filter_()
library("lazyeval") ## for interp()
colname <- "speed"
filter_(cars,interp(~ var == 4, var = as.name(colname)))
source to share