[R] Testing for membership in an array of strings

[Dirk Eddelbuettel]

On 2 July 2009 at 09:13, Neil Tiffin wrote:
| Given a array of string values, for example lets say "mary", "bob",  
| "danny", "sue", and "jane".
| 
| I am trying to determine how to perform a logical test to determine if  
| a variable is an exact match for one of the string values in the array  
| when the number of strings in the array is variable and without using  
| a for loop and without comparing each value.  Considering the power of  
| R, I thought this would be easy, but its not obvious to me.
| 
| Now I may not yet be one with the R fu so a bit more context.
| 
| I have a data frame that contains a column with text values. What I am  
| trying to do is use the subset function on the data frame to select  
| only data for "sue" or "jane" (for example.)  But maybe I have not  
| taken the correct approach?
| 
| So obviously I could do something like the following.
| 
| subset( data_frame, name = "sue" | name == "jane", select = c(name,  
| age, birthdate))
| 
| However, my subset needs to be much more than 2 and being lazy I do  
| not want to type "| name == "some text" for each one.
| 
| Is there an other way?

Yup, e.g. using the %in% operator:

> set.seed(42)  # fix rng so that you get the same data.frame
> neil <- data.frame(rownb=1:20, name=sample(c("mary", "bob", "danny", "sue",
> "jane"), 20, replace=TRUE))
> head(neil)    # quick sanity check
  rownb  name
1     1  jane
2     2  jane
3     3   bob
4     4  jane
5     5   sue
6     6 danny
> neil[ neil$name %in% c("sue", "jane"), ]
   rownb name
1      1 jane
2      2 jane
4      4 jane
5      5  sue
7      7  sue
9      9  sue
10    10  sue
12    12  sue
13    13 jane
16    16 jane
17    17 jane
> 

Cheers, Dirk

-- 
Three out of two people have difficulties with fractions.