Range on a field containing NAs

问题:

I'm using a data set where the 11th column on a csv file has numeric data. It contains some NA values too. Here is the str of the object:

str(dataheart)
 num [1:4706] 14.3 18.5 18.1 NA NA NA 17.7 18 15.9 NA ...

So, as a new student of R, I had expected the result of range(dataheart) to be the min and max values.From looking at the CSV file with data, I know that the min and max are 10.1 and 21.9.

But the above returns a vector

[1] NA NA

Is my understanding of this function incorrect?

回答1:

You need

range(x,na.rm=TRUE)

see ?range

For extra credit, here's a list of the functions in the base and stats packages that use na.rm:

uses_na_rm <- function(x) is.function(fx <- get(x)) && 
                         "na.rm" %in% names(formals(fx))
basevals <- ls(pos="package:base")
basevals[sapply(basevals,uses_na_rm)]
##  [1] "colMeans"                "colSums"                
##  [3] "is.unsorted"             "mean.default"           
##  [5] "pmax"                    "pmax.int"               
##  [7] "pmin"                    "pmin.int"               
##  [9] "range.default"           "rowMeans"               
## [11] "rowsum.data.frame"       "rowsum.default"         
## [13] "rowSums"                 "Summary.data.frame"     
## [15] "Summary.Date"            "Summary.difftime"       
## [17] "Summary.factor"          "Summary.numeric_version"
## [19] "Summary.ordered"         "Summary.POSIXct"        
## [21] "Summary.POSIXlt"        

statvals <- ls(pos="package:stats")
statvals[sapply(statvals,uses_na_rm)]
## [1] "density.default"  "fivenum"          "heatmap"          "IQR"             
## [5] "mad"              "median"           "median.default"   "medpolish"       
## [9] "quantile.default" "sd"               "var"   

For further consideration of which functions in R deal with NAs and how, one could do an analogous search for functions with an na.action argument (lm and friends).