variable value occuring on 2 dates R - r

Question

I want to find who had an apple or an orange on at least 2 different (unique) dates. I would like to create a new column with a binary indicator for whether an individual had an orange or an apple on at least two dates (1=yes, 0=no).
The nearest I've come is this plyr code.
df1<- ddply(df, .(names, fruit), mutate, acne = ifelse(fruit=="apple" | fruit=="orange" & length(unique(dates))>=2,1,0))
This is not the solution however. anne gets apples twice but on the same date, so she should not get a 1 here. Similarly ted gets a 1, even though he only got an apple once.
This is closer, but still not correct. It gives a 1 to any fruit that has occurred twice. Need the fruit to occur twice per person on two individual dates per person
df2<- ddply(df, .(fruit), mutate, acne = ifelse(length(unique(dates))>=2, 1, 0
##this one gives a 1 to any fruit that has occurred twice. Need the fruit to occur twice per person on two individual dates per person.
If anyone could point me in the right direction here I would be very grateful.
Thank you in advance
SAMPLE DF
names<-as.character(c("john", "john", "philip", "ted", "john", "john", "anne", "john", "mary","anne", "mary","mary","philip","mary", "su","mary", "jim", "sylvia", "mary", "ted","ted","mary", "sylvia", "jim", "ted", "john", "ted"))
dates<-as.Date(c("2010-07-01", "2010-07-13", "2010-05-12","2010-02-14","2010-06-30","2010-08-15", "2010-03-21","2010-04-04","2010-09-01", "2010-03-21", "2010-12-01", "2011-01-01", "2010-08-12", "2010-11-11", "2010-05-12", "2010-12-03", "2010-07-12", "2010-12-21", "2010-02-18", "2010-10-29", "2010-08-13", "2010-11-11", "2010-05-12", "2010-04-01", "2010-05-06", "2010-09-28", "2010-11-28" ))
fruit<-as.character(c("kiwi","apple","mango", "banana","strawberry","orange","apple","raspberry", "orange","apple","orange", "apple", "strawberry", "apple", "pineapple", "peach", "orange", "nectarine", "grape","banana", "melon", "apricot", "plum", "lychee", "mango", "watermelon", "apple" ))
df<-data.frame(names,dates,fruit)
df
Desired ouput
names dates fruit v1
7 anne 2010-03-21 apple 0
10 anne 2010-03-21 apple 0
17 jim 2010-07-12 orange 0
24 jim 2010-04-01 lychee 0
1 john 2010-07-01 kiwi 1
2 john 2010-07-13 apple 1
5 john 2010-06-30 strawberry 1
6 john 2010-08-15 orange 1
8 john 2010-04-04 raspberry 1
26 john 2010-09-28 watermelon 1
9 mary 2010-09-01 orange 1
11 mary 2010-12-01 orange 1
12 mary 2011-01-01 apple 1
14 mary 2010-11-11 apple 1
16 mary 2010-12-03 peach 1
19 mary 2010-02-18 grape 1
22 mary 2010-11-11 apricot 1
3 philip 2010-05-12 mango 0
13 philip 2010-08-12 strawberry 0
15 su 2010-05-12 pineapple 0
18 sylvia 2010-12-21 nectarine 0
23 sylvia 2010-05-12 plum 0
4 ted 2010-02-14 banana 0
20 ted 2010-10-29 banana 0
21 ted 2010-08-13 melon 0
25 ted 2010-05-06 mango 0
27 ted 2010-11-28 apple 0

Answer 1

this should probably do the trick:
v1 = ave(1:nrow(df),df$names,FUN=function(x) length(unique(df$dates[x[df$fruit[x]
%in% c("orange","apple")]]))>1)
df$v1 = v1
df = df[order(df$names),]

Answer 2

If I understood correctly, for the purpose of your problem, apples == oranges. So the plan is
to (1) create a small data.frame where fruits are oranges or apples only, as you don't care about other fruits, (b) select only unique date/name rows, (c) aggregate by name and (d) merge back to your original data.frame to get your result:
ndf <- subset(df, fruit %in% c("apple", "orange"))
ndf <- ndf[!duplicated(ndf[, c("names", "dates")]), ]
Here you can use table, but I prefer aggregate
v <- aggregate(rep(1, nrow(ndf)), by = ndf[, "names", drop = FALSE], sum)
v$x <- ifelse(v$x > 1, 1, 0)
rv <- merge(df, v)
It is a bit longer, codewise, than other answers but clear and most certainly does the job.
You could just use aggregate without the first two parts, but if you have huge data.frame, with lots of names aggregating for every name can prove very costly.

Answer 3

I did something similar to #amit's solution using by. Rownames got mangled during do.call, but you can fix that.
result <- by(df, INDICES = df$names, FUN = function(x) {
if (length(unique(x$dates)) == 1) {
x$index <- 0
return(x)
}
ao.sum <- sum(x$fruit %in% c("apple", "orange"))
if (ao.sum < 2) x$index <- 0 else x$index <- 1
x
})
do.call("rbind", result)
names dates fruit index
anne.7 anne 2010-03-21 apple 0
anne.10 anne 2010-03-21 apple 0
jim.17 jim 2010-07-12 orange 0
jim.24 jim 2010-04-01 lychee 0
john.1 john 2010-07-01 kiwi 1
john.2 john 2010-07-13 apple 1
john.5 john 2010-06-30 strawberry 1
john.6 john 2010-08-15 orange 1
john.8 john 2010-04-04 raspberry 1
john.26 john 2010-09-28 watermelon 1
mary.9 mary 2010-09-01 orange 1
mary.11 mary 2010-12-01 orange 1
mary.12 mary 2011-01-01 apple 1
mary.14 mary 2010-11-11 apple 1
mary.16 mary 2010-12-03 peach 1
mary.19 mary 2010-02-18 grape 1
mary.22 mary 2010-11-11 apricot 1
philip.3 philip 2010-05-12 mango 0
philip.13 philip 2010-08-12 strawberry 0
su su 2010-05-12 pineapple 0
sylvia.18 sylvia 2010-12-21 nectarine 0
sylvia.23 sylvia 2010-05-12 plum 0
ted.4 ted 2010-02-14 banana 0
ted.20 ted 2010-10-29 banana 0
ted.21 ted 2010-08-13 melon 0
ted.25 ted 2010-05-06 mango 0
ted.27 ted 2010-11-28 apple 0