如何使用循环替换基于 r 数据框中另一列的平均值的值

Question

If I have the following data frame in r:

Pitcher Pitch.Spin..rpm. 
A     2350
A     2400
A     2233
A     1100
B     2145
B     2200
B     2340
B     1050

and I wanted to write a loop in R to replace the low values in A and B with their respective means that excluded the bad readings so that the output would be:

A     2350
A     2400
A     2233
A     2328
B     2145
B     2200
B     2340
B     2228

How would I go about doing that? Below was my attempt and my issue comes from not being sure how to properly reference the Pitcher value in the specific row

for (i in 1:nrow(data)){
  if (data$Pitch.Spin..rpm. < 1500)
  data$Pitch.Spin..rpm. <- mean(data$Pitch.Spin..rpm.[Pitcher == {i}],na.rm = TRUE)
}

Answer 1

We could do this with a group by operation. After grouping by 'Pitcher', mutate the 'Pitch.Spin..rpm.' by replace ing the elements that are less than 1500 with the the mean of that column

library(dplyr)
data <- data %>%
   group_by(Pitcher) %>%
   mutate(`Pitch.Spin..rpm.` = replace(`Pitch.Spin..rpm.`, 
        `Pitch.Spin..rpm.` < 1500, mean(`Pitch.Spin..rpm.`, na.rm = TRUE)))

Answer 2

A base R solution, with ave .

ave(df$`Pitch.Spin..rpm.`, df$Pitcher, FUN = function(x){
  i <- x < 1500
  if(any(i)) x[i] <- mean(x[!i])
  x
})
#[1] 2350.000 2400.000 2233.000 2327.667 2145.000 2200.000 2340.000
#[8] 2228.333

Now assign this result back to the df's column.

df$Pitch.Spin..rpm. <- ave(df$Pitch.Spin..rpm., df$Pitcher, FUN = function(x){
  i <- x < 1500
  if(any(i)) x[i] <- mean(x[!i])
  x
})

df
#  Pitcher Pitch.Spin..rpm.
#1       A         2350.000
#2       A         2400.000
#3       A         2233.000
#4       A         2327.667
#5       B         2145.000
#6       B         2200.000
#7       B         2340.000
#8       B         2228.333

Answer 3

An approach using dplyr and ifelse() to replace values is next:

library(dplyr)

#Data
df <- structure(list(Pitcher = c("A", "A", "A", "A", "B", "B", "B", 
"B"), Pitch.Spin..rpm. = c(2350L, 2400L, 2233L, 1100L, 2145L, 
2200L, 2340L, 1050L)), class = "data.frame", row.names = c(NA, 
-8L))

Code:

#Code
df %>% group_by(Pitcher) %>%
  mutate(NewVar=ifelse(Pitch.Spin..rpm.<1500,NA,Pitch.Spin..rpm.),
         Mean=mean(NewVar,na.rm=T),
         Pitch.Spin..rpm. = ifelse(is.na(NewVar),Mean,Pitch.Spin..rpm.)) %>%
  select(-c(NewVar,Mean))

Output:

# A tibble: 8 x 2
# Groups:   Pitcher [2]
  Pitcher Pitch.Spin..rpm.
  <chr>              <dbl>
1 A                  2350 
2 A                  2400 
3 A                  2233 
4 A                  2328.
5 B                  2145 
6 B                  2200 
7 B                  2340 
8 B                  2228.

A way to do with loop is next but you have to save results in a list:

#Unique pitcher
val <- unique(df$Pitcher)
#Create empty list
List <- list()
#Loop
for(i in val)
{
  #Isolate data
  data1 <- subset(df,Pitcher==i)
  #Compute mean
  meanval <- mean(data1$Pitch.Spin..rpm.[!data1$Pitch.Spin..rpm.<1500])
  #Replace
  data1$Pitch.Spin..rpm.[data1$Pitch.Spin..rpm.<1500]<-meanval
  #Save in list
  List[[i]] <- data1
}
#Now bind the list
newdf <- do.call(rbind,List)
rownames(newdf) <- NULL

Output:

  Pitcher Pitch.Spin..rpm.
1       A         2350.000
2       A         2400.000
3       A         2233.000
4       A         2327.667
5       B         2145.000
6       B         2200.000
7       B         2340.000
8       B         2228.333