CODEWITHSUNDEEP
rfor-loopdataframe
user2325155
user2325155

Reputation: 147

for loop through data frame and looping with unique values

I'm trying to work on code to build a function for three stage cluster sampling, however, I am just working with dummy data right now so I can understand what is going into my function.

I am working on for loops and have a data frame with grouped values. I'm have a data frame that has data:

    Cluster group   value   value.K.bar  value.M.bar      N.bar
1       1       A   1        1.5            2.5             4
2       1       A   2        1.5            2.5             4
3       1       B   3        4.0            2.5             4
4       1       B   4        4.0            2.5             4
5       2       B   5        4.0            6.0             4
6       2       C   6        6.5            6.0             4
7       2       C   7        6.5            6.0             4

and I am trying to run the for loop 

n <- dim(data)[1]
e <- 0
total <- 0
for(i in 1:n) {e = data.y$value.M.bar[i] - data$N.bar[i]
               total = total + e^2}

My question is: Is there a way to run the same loop but for the unique values in the group? Say by:

Group 'A', 'B', 'C'

Any help would be greatly appreciated!

Edit: for correct language

Upvotes: 2

Views: 6242

Answers (1)

agstudy
agstudy

Reputation: 121568

You can use by for example, to apply your data per group. First I wrap your code in a function that take data as input.

get.total <- function(data){
  n <- dim(data)[1]
  e <- 0
  total <- 0
  for(i in 1:n) {
    e <- data$value.M.bar[i] - data$N.bar[i] ## I correct this line
    total <- total + e^2
  } 
  total
}

Then to compute total just for group B and C you do this :

    by(data,data$group,FUN=get.total)
data$group: A
[1] 4.5
---------------------------------------------------------------------------------------------------- 
data$group: B
[1] 8.5
---------------------------------------------------------------------------------------------------- 
data$group: C
[1] 8

But better , Here a vectorized version of your function

 by(data,data$group,
        function(dat)with(dat, sum((value.M.bar - N.bar)^2)))

Upvotes: 4

Related Questions