tags:

views:

73

answers:

2

(second question today - must be a bad day)

I have a dataframe with various columns, inculding a concentration column (numeric), a flag highlighting invalid results (boolean) and a description of the problem (character)

df <- structure(list(x = c(1, 2, 3, 4, 5, 6, 7, 8, 9, 10), rawconc = c(77.4, 
52.6, 86.5, 44.5, 167, 16.2, 59.3, 123, 1.95, 181), reason = structure(c(NA, 
NA, 2L, NA, NA, NA, 2L, 1L, NA, NA), .Label = c("Fails Acceptance Criteria", 
"Poor Injection"), class = "factor"), flag = c("False", "False", 
"True", "False", "False", "False", "True", "True", "False", "False"
)), .Names = c("x", "rawconc", "reason", "flag"), row.names = c(NA, 
-10L), class = "data.frame")

I can create a column with the numeric level of the reason column

df$level<-as.numeric(df$reason)
df
    x rawconc                    reason  flag level
1   1   77.40                      <NA> False    NA
2   2   52.60                      <NA> False    NA
3   3   86.50            Poor Injection  True     2
4   4   44.50                      <NA> False    NA
5   5  167.00                      <NA> False    NA
6   6   16.20                      <NA> False    NA
7   7   59.30            Poor Injection  True     2
8   8  123.00 Fails Acceptance Criteria  True     1
9   9    1.95                      <NA> False    NA
10 10  181.00                      <NA> False    NA

and here's what I want to do to create a column with 'level' many stars, but it fails

df$stars<-paste(rep("*",df$level)sep="",collapse="")
Error: unexpected symbol in "df$stars<-paste(rep("*",df$level)sep"

df$stars<-paste(rep("*",df$level),sep="",collapse="")
Error in rep("*", df$level) : invalid 'times' argument

rep("*",df$level)
Error in rep("*", df$level) : invalid 'times' argument

df$stars<-paste(rep("*",pmax(df$level,0,na.rm=TRUE)),sep="",collapse="")
Error in rep("*", pmax(df$level, 0, na.rm = TRUE)) : 
  invalid 'times' argument

It seems that rep needs to be fed one value at a time. I feel that this should be possible (and my gut says 'use lapply' but my apply fu is v. poor)

Any one want to try ?

+2  A: 

I think that you will need an apply-type function. This will work:

df[is.na(df$level),"level"] <- 0
df$level <- sapply(df$level, function(x) paste(rep("*",x),collapse=""))

You would be better using sapply than lapply in this instance since it returns a vector instead of a list.

From the help for rep:

If 'times' consists of a single integer, the result consists of the whole input repeated this many times. If 'times' is a vector of the same length as 'x' (after replication by 'each'), the result consists of 'x[1]' repeated 'times[1]' times, 'x[2]' repeated 'times[2]' times and so on.

One problem with using rep with a vector for the times parameter is that it just returns a vector and it discards instances when times=0. You can see this with this command: rep(rep("*", nrow(df)), times=df$level).

Shane
+3  A: 

You could create stars vector as

vstars <- sapply(1L:nlevels(df$reason), function(i) paste(rep("*",i),collapse=""))
vstars
# [1] "*"  "**"

And then indexing it with df$reason (which works because its a factor):

vstars[df$reason]
# [1] NA   NA   "**" NA   NA   NA   "**" "*"  NA   NA

For large data.frame should be much faster then paste in each row.

Marek