Skip to content

Subsetting a data frame

2 messages · natalie.vanzuydam, jim holtman

natalie.vanzuydam
# 
Hi R users,

I really need help with subsetting  data frames:

I have a large database of medical records and I want to be able to match
patterns from a list of search terms .

I've used this simplified data frame in a previous example:


db <- structure(list(ind = c("ind1", "ind2", "ind3", "ind4"), test1 = c(1, 
2, 1.3, 3), test2 = c(56L, 27L, 58L, 2L), test3 = c(1.1, 28, 
9, 1.2)), .Names = c("ind", "test1", "test2", "test3"), class =
"data.frame", row.names = c(NA, 
-4L)) 

terms_include <- c("1","2","3") 
terms_exclude <- c("1.1","1.2","1.3") 


So in this example I want to include all the terms from terms include as
long as they don't occur with terms exclude in the same row of the data
frame.

Previously I was given this function which works very well if you want to
match exactly:


f <- function(x)  !any(x %in% terms_exclude) && any(x %in% terms_include) 
db[apply(db[, -1], 1, f), ] 

   ind test1 test2 test3 
2 ind2     2    27  28.0 
4 ind4     3     2   1.2 


I would like to know if there is a way to write a similar function that
looks for matches that start with the query string:  as in
grepl("^pattern",x)  

I started writing a function but am not sure how to get it to return the
dataframe or matrix:


for (i in 1:length(terms_include)){
db_new <- apply(db,2, grepl,pattern=i)
}

Applying this function gives me:

db_new <- structure(c(FALSE, FALSE, TRUE, FALSE, FALSE, FALSE, TRUE, TRUE, 
FALSE, FALSE, FALSE, FALSE, FALSE, FALSE, FALSE, FALSE), .Dim = c(4L, 
4L), .Dimnames = list(NULL, c("ind", "test1", "test2", "test3"
)))

So the above is searching the pattern anywhere in the dataframe instead of
just at the beginning of the string.  

How would I incorporate look for terms to include but don't return the row
of the data frame if it also includes one of the terms to exclude while
using partial matching?

I hope that this makes sense.

Many thanks,
Natalie

-----
Natalie Van Zuydam

PhD Student
University of Dundee
nvanzuydam at dundee.ac.uk
--
View this message in context: http://r.789695.n4.nabble.com/Subsetting-a-data-frame-tp4160127p4160127.html
Sent from the R help mailing list archive at Nabble.com.
jim holtman
# 
does this do what you want:
+ 2, 1.3, 3), test2 = c(56L, 27L, 58L, 2L), test3 = c(1.1, 28,
+ 9, 1.2)), .Names = c("ind", "test1", "test2", "test3"), class =
+ "data.frame", row.names = c(NA,
+ -4L))
+     pat <- paste("^(", paste(inc, collapse = "|"), ")", sep = '')
+     patex <- paste(exc, collapse = "|")
+     isMatch <- apply(obj, 1, function(x) any(grepl(pat, x)))
+     notMatch <- !apply(obj, 1, function(x) any(grepl(patex, x)))
+     obj[isMatch & notMatch,]
+ }
ind test1 test2 test3
1 ind1   1.0    56   1.1
2 ind2   2.0    27  28.0
3 ind3   1.3    58   9.0
4 ind4   3.0     2   1.2
ind test1 test2 test3
2 ind2     2    27    28

        
On Mon, Dec 5, 2011 at 6:32 AM, natalie.vanzuydam <nvanzuydam at gmail.com> wrote: