R max of multiple categories [duplicate] - r

Question

This question already has answers here:
Aggregate a dataframe on a given column and display another column
(8 answers)
Closed 5 years ago.
I've got data somewhat like this (of course with many more rows):
Age Work Zone SomeNumber
26 1 2.61
32 4 8.42
41 2 9.71
45 2 4.14
64 3 6.04
56 1 5.28
37 4 7.93
I want to get the maximum SomeNumber for each zone at or below each age. SomeNumber increases with age, so I expect that the highest SomeNumber in Zone 2 by an under-32-y/o is by a guy who's age 31, but it could in fact be a guy age 27.
To do this I've written a nested for loop:
for(i in zonelist){
temp = data[data$zone==i,]
for(j in 1:max(data$age)){
temp.lessequal=c(temp.lessequal,max((temp[temp$Age<=j,])$SomeNumber))
}
#plot temp.lessequal or save it at this point
}
which of course is tremendously slow. How can I do this faster? I've looked at the order function to sort by two columns at once, but that doesn't let me take the max of each group.

Answer 1

Data:
df1 <- read.table(text='Age Work_Zone SomeNumber
26 1 2.61
32 4 8.42
41 2 9.71
45 2 4.14
64 3 6.04
56 1 5.28
37 4 7.93',
header = TRUE)
Code:
df2 <- with( df1, df1[ Age <= 32, ] ) # extract rows with Age <= 32
# get maximum of someNumber by aggregating with work_zone and then merging with df1 to combine the age column
merge(aggregate(SomeNumber ~ Work_Zone, data = df2, max), df2)
# Work_Zone SomeNumber Age
# 1 1 2.61 26
# 2 4 8.42 32

Answer 2

It seems OP is looking for max value based on <= condition on a particular column (Age).
The use of sqldf comes very handy in such cases in order to explain the logic. One solution could be:
# Data
df <- read.table(text = "Age Work_Zone SomeNumber
26 1 2.61
32 4 8.42
41 2 9.71
45 2 4.14
64 3 6.04
56 1 5.28
37 4 7.93", header = T, stringsAsFactors = F)
library(sqldf)
df3 <- sqldf("select df1.Work_Zone, df1.Age, max(df2.SomeNumber) from df df1
inner join df df2 on df1.Work_Zone = df2.Work_Zone
WHERE df2.Age <= df1.Age
GROUP BY df1.Work_Zone, df1.Age")
# Result:
# Work_Zone Age max(df2.SomeNumber)
# 1 1 26 2.61
# 2 1 56 5.28
# 3 2 41 9.71
# 4 2 45 9.71
# 5 3 64 6.04
# 6 4 32 8.42
# 7 4 37 8.42

Answer 3

Using the library data.table you can select the rows which are less than required age, then output the max(somenumber) and their respective age for each Workzone ie group by workzone.
library(data.table)
setDT(df1)[Age<=32,.(max(SomeNumber),Age),by=Work_Zone]
Work_Zone V1 Age
1: 1 2.61 26
2: 4 8.42 32