How to subset times in R - r

Question

In R I have data
USER BIRTH
11 "2013-01-11 22:31:11"
121 "2014-12-26 04:07:35"
...
I want to create a new data set data_new that contain all USER in the time 10 o'clock to 11 o'clock.
The types of USER and BIRTH are strings/characters. I tried this:
data_new= data$BIRTH > as.POSIXct("10:00:00", format="%H:%M:%S")
& data$BIRTH < as.POSIXct("11:00:00", format="%H:%M:%S")
but here R gives we FALSE for all entries, so this don't work.
How can I solve this?
Update
Say I want to find the number of users for all hours. I use the answer and try this
u=c()
for(j in 1:24) {
data_new=data[times > "00:00:00"+(j-1) & times < "01:00:00"+j ,]
#saving the number of users in vector u
u[j]=dim(data_new)[1]
}
but R can't figure out the term "00:00:00"+(j-1).

Answer 1

If df is your data frame:
df <- read.table(text = 'USER BIRTH
11 "2013-01-11 22:31:11"
121 "2014-12-26 04:07:35"
121 "2014-12-26 10:07:35"
121 "2014-12-26 11:07:35"
121 "2014-12-26 10:38:35"', header = T)
df$BIRTH <- ymd_hms(df$BIRTH)
times <- strftime(df$BIRTH, format = "%H:%M:%S")
df[times > "10:00:00" & times < "11:00:00",]
Output:
USER BIRTH
3 121 2014-12-26 10:07:35
5 121 2014-12-26 10:38:35

Answer 2

One way to do something to each subset of your data is to use the split-lapply paradigm. In this case, you would convert data$BIRTH to POSIXlt and split by the hour component of the POSIXlt object. That will give you a list where each list element contains all the data for a specific hour.
data <- read.csv(text = "USER,BIRTH
11,2013-01-11 22:31:11
12,2014-12-26 04:07:35
21,2014-12-26 10:07:35
121,2014-12-26 11:07:35
112,2014-12-26 10:38:35")
data_by_hour <- split(data, as.POSIXlt(data$BIRTH)$hour)
Then you can use lapply (or sapply) to do whatever you want to each of those subsets. To count the number of observations per hour:
# number of observations for each hour
sapply(data_by_hour, nrow)
4 10 11 22
1 2 1 1
You can also do this with xts.
library(xts)
# Create xts object from 'data' data.frame
# Note: xts objects are based on a matrix, so you cannot have columns with
# mixed types like you can with a data.frame.
x <- xts(data["USER"], as.POSIXct(data$BIRTH))
period.apply(x, endpoints(x, "hours"), nrow)
# USER
# 2013-01-11 22:31:11 1
# 2014-12-26 04:07:35 1
# 2014-12-26 10:38:35 2
# 2014-12-26 11:07:35 1
Note that you can do time-of-day subsetting with xts. It avoids potential locale-related collation order issues caused by using logical operators on character strings.
x["T10:00/T11:00"]
# USER
# 2014-12-26 10:07:35 21
# 2014-12-26 10:38:35 112