R writing a function to avoid for loop - r

Question

Hi I am trying to learn ways in which I can avoid loops in my codes.
I have an example data here:
options(warn=-1) #Turning warnings off here
Company=c("A","C","B","B","A","C","C","A","B","C","B","A")
CityID=as.character(c(1,1,1,2,2,2,3,3,3,4,4,4))
Value=c(120.5,123,125,122.5,122.1,121.7,123.2,123.7,120.7,122.3,120.1,122)
Sales=c(1,1,0,0,0,1,1,0,1,0,1,0)
df=data.frame(Company,CityID,Sales,Value)
df$new_value=0
I also created a custom function (simple example only for testing purposes) as below.
funcCity12 = function(data){
data_new=data[which(data$CityID == '1'|data$CityID == '2'),]
for (i in 1:nrow(data_new)){
data_company=df[(df$Company)==data_new[i,'Company'] & !df$CityID==1 & !df$CityID==2,]
data_new[i,'new_value'] = max(data_company[data_company$Sales==1,]$Value) #Note we take the maximum value here
}
data_new
}
df2=funcCity12(data=df) # obtaining the result here
Now I am trying to write a function to avoid the for loop in the previous function.
funcCity12_no_loop = function(x,df){
data_company=df[(df$Company)==x[,'Company'] & !df$CityID==1 & !df$CityID==2,]
x[,'new_value'] = max(data_company[data_company$Sales==1,]$Value) #Note we take the maximum value here
x
}
funcCity12_no_loop(x=df[1,],df=df) #Output for the first row of df1
This seems to be working when I input the rows individually. What I am stuck at is how to run this function for all rows of the dataframe. I am not sure if the 2nd function requires more changes for this purpose. Any help is appreciated. Thanks in advance.
P.S. For the second function, my initial reaction was to create a for loop and loop through the observations, but that defeats the whole purpose.
EDIT
This is based on #eonurk's answer
zz=apply(df,1, function(x){
data_company=df[(df$Company)==x[1] & !df$CityID==1 & !df$CityID==2,]
x[5] = max(data_company[data_company$Sales==1,]$Value) #Note we take the maximum value here
x
})
Output is shown below:

Answer 1

You can use apply function to reach out each individual observation of your dataframe.
For instance, you can multiplicate Values and Sales columns for no reason at all with following:
apply(df,1, function(x){ as.numeric(x["Sales"])*as.numeric(x["Value"])})
Edit:
Now you just need to use dplyr package
zz=apply(df,1, function(x){
data_company=df[(df$Company)==x[1] & !df$CityID==1 & !df$CityID==2,]
x[5] = max(data_company[data_company$Sales==1,]$Value) #Note we take the maximum value here
x
}) %>% as.data.frame %>% t

Answer 2

Here is one way without a loop. First we filter based on your criteria, then we group by company and calculate the max, then we join the dataframe to the original dataset (also filtered based on your criteria). I didn't make it a function, but the building blocks are all there.
library(tidyverse)
list(
df %>%
filter(CityID %in% 1:2) %>%
select(-new_value),
df %>%
filter(! CityID %in% 1:2 & Sales == 1) %>%
group_by(Company) %>%
summarise(new_value = max(Value))
) %>%
reduce(full_join, by = "Company")
#> Company CityID Sales Value new_value
#> 1 A 1 1 120.5 NA
#> 2 C 1 1 123.0 123.2
#> 3 B 1 0 125.0 120.7
#> 4 B 2 0 122.5 120.7
#> 5 A 2 0 122.1 NA
#> 6 C 2 1 121.7 123.2