Replace specific column "words" into number or blank

Question

Input table

Patients  Hospital   Drug   Response
1         AAA        a      Good
1         AAA        a      Bad
2         BBB        a      Bad
3         CCC        b      Good
4         CCC        c      Bad
5         DDD        e      undefined

Output file

Patients  Hospital   Drug   Response
1         AAA        a      1
1         AAA        a      -1
2         BBB        a      -1
3         CCC        b      1
4         CCC        c      -1
5         DDD        e

How to replace 3 texts in one column to number and blank?

"good in Reponse column" to "1" "bad in Reponse column" to "-1" "undefined in Reponse column" to " "

Data:

structure(list(Patients = c(1L, 1L, 2L, 3L, 4L, 5L), Hospital = structure(c(1L, 
1L, 2L, 3L, 3L, 4L), .Label = c("AAA", "BBB", "CCC", "DDD"), class = "factor"), 
    Drug = structure(c(1L, 1L, 1L, 2L, 3L, 4L), .Label = c("a", 
    "b", "c", "e"), class = "factor"), Response = structure(c(2L, 
    1L, 1L, 2L, 1L, 3L), .Label = c("Bad", "Good", "undefined"
    ), class = "factor")), .Names = c("Patients", "Hospital", 
"Drug", "Response"), class = "data.frame", row.names = c(NA, 
-6L))

I think Gavin dputted the sample data. – Roman Luštrik Apr 11 '11 at 05:38 — Roman Luštrik, Apr 11 '11 at 05:38

score 17 · Answer 1 · answered Apr 10 '11 at 16:09

17

You can do this with one line by changing the labels of the factor Response:

> within(df, Response <- factor(Response, labels = c(-1, 1, "")))
  Patients Hospital Drug Response
1        1      AAA    a        1
2        1      AAA    a       -1
3        2      BBB    a       -1
4        3      CCC    b        1
5        4      CCC    c       -1
6        5      DDD    e

answered Apr 10 '11 at 16:09

Gavin Simpson

170,508
25
396
453

1

+1 nice! Of course, `Response` has to be a factor (which it probably is). – csgillespie Apr 10 '11 at 16:25
1

alternatively, `sapply(Data$Response,switch,'Good'=1,'Bad'=-1,'undefined'="")` is a more general way of doing that, but it is definitely slower. Plus, when using switch, Data$Response has to be a character vector, otherwise you get faulty results. – Joris Meys Apr 10 '11 at 21:09
How do you do this without overwriting the original data frame? – stackoverflowuser2010 Feb 27 '14 at 00:03
@stackoverflowuser2010 (So *now* you want my R advice? :-) If you want to create the vector named `Response` in the output from `within()`, simply `factor(df$Response, labels = c(-1, 1, ""))` will do it. If you want a new data frame then just assign the result of `within()` to a new object `df2 <- within(df, Response <- factor(Response, labels = c(-1, 1, "")))`. `within()` (and the similar `transform()`) are convenience functions for interactive use they just return a modified data frame so assigning the result to something other than the data frame used will result in a new data frame. – Gavin Simpson Feb 27 '14 at 02:01
@GavinSimpson Im a litte confused with your answer of creating a vector named Response in the output. I want a new column(vector) in my dataframe that is now coded.`within(df, obs <- factor(df$Response, labels = c(1,0,-1)))` it does NOT add a new perm. column. it produces a temporary list. so you still have to write a new object or write over the old object. Or am I missing something? – Kerry Apr 01 '14 at 18:23
@Kerry `within` returns a modified data frame, you need to assign it, either to a new object or to the existing object name. This is R - most things don't get modified without assigning them. Also note that you don't need `factor(df$Response, ....)` but just `factor(Response, ....)` because `Response` is made visible by `within`. – Gavin Simpson Apr 01 '14 at 23:06

score 5 · Answer 2 · edited May 23 '17 at 11:48

Catherine, your questions could still be answered by a very basic textbook in R. Please see Dirk's comment in your previous question.

Answer

If d is your data frame, then:

d[d$Response == "Good",]$Response = 1
d[d$Response == "Bad",]$Response = -1
d[d$Response == "undefined",]$Response = ""

I'm guessing (I may be wrong) that "Undefined" is missing data. In which case, use NA rather than a blank. Any basic R book will describe NA's

score 3 · Answer 3 · answered Apr 10 '11 at 16:02

3

If your data is in a data frame df

df$Response[df$Response == "Good"] <- 1
df$Response[df$Response == "Bad"] <- -1
df$Response[df$Response == "undefined"] <- ""

answered Apr 10 '11 at 16:02

Noah

2,574
1
18
12

Roman Luštrik · Answer 4 · 2011-04-10T19:08:38.433

2

You can use a simple ifelse() statement.

cath <- data.frame(nmbrs = runif(10), words = sample(c("good", "bad"), 10, replace = TRUE))
cath$words <- ifelse(cath$words == "good", 1, ifelse(cath$words == "bad", -1, ""))

edited Apr 10 '11 at 19:08

answered Apr 10 '11 at 16:05

Roman Luštrik

69,533
24
154
197

Replace specific column "words" into number or blank

4 Answers4

Linked