CODEWITHSUNDEEP
rdataframemeancalculated-columns
KKL234
KKL234

Reputation: 367

R calculate new column from second data frame

I have 2 data frames that I'm trying to keep separate, but would like to calculate a new column in df1 based on the means of rows in df2:

df1: 

Name <- c("t1", "t2", "t3")    
Diam <- c(95, 104, 72)    
Rad <- c(47.5, 52, 36)    
df1 <- data.frame(Name, Diam, Rad)

df2: 

Name <- c("t1A", "t1B", "t2A", "t2B", "t2C", "t3B")    
Width <- c(4000, 3200, 2300, 2700, 2100, 3500)    
df2 <- data.frame(Name, Width)

I'd like to add a fourth column to df1 that calculates the means of "Width" in df2 when the first 2 characters of "Name" are identical. My intended output looks like:

df3: 

Name <- c("t1", "t2", "t3")    
Diam <- c(95, 104, 72)     
Rad <- c(47.5, 52, 36)    
WidthMean <- c(3600, 2366.667, 3500)    
df3 <- data.frame(Name, Diam, Rad, WidthMean)

I can do it by splitting the "Name" column in df2, then finding the mean according to the new split name, then adding the mean values from df2 as a new column to df1. However, I have to do this to approximately 75 data.frames and was hoping to find a faster way to process the data (and with fewer lines of code!).

Current code: 

df2$"NameID" <- df2$Name    
df2$NameID <- gsub(".?$", "", df2$NameID)    
df2out <- aggregate(Width ~ NameID, data = df2, FUN = mean)    
df1$"WidthMean" <- df2out$Width    
df1

Upvotes: 2

Views: 2133

Answers (3)

maloneypatr
maloneypatr

Reputation: 3622

Here's a solution using dplyr and a loop to read in & merge your files

library(dplyr)
library(gdata)

setwd('TO_THE_FOLDER_WHERE_ALL_YOUR_FILES_ARE')

names <- list.files()
names <- sub('dbh.csv', '', names)
names <- sub('bai.csv', '', names)
uniqueNames <- unique(names)

for(uniqueName in uniqueNames){

  df1_name <- paste0(uniqueName, 'dbh.csv')
  df1 <- read.csv(df1_name)
  df2_name <- paste0(uniqueName, 'bai.csv')
  df2 <- read.csv(df2_name)

  df2 %>%
    mutate(partialName = str_sub(Name, 1, 2)) %>%
    group_by(partialName) %>%
    dplyr::summarise(meanWidth = mean(Width)) %>%
    merge(df1, by.x = 'partialName', by.y = 'Name') %>%
    select(partialName, Diam, Rad, meanWidth) %>%
    rename.vars('partialName', 'Name')

  assign(df2, uniqueName)
  #write.csv(df2, paste0(uniqueName, '.csv'), row.names = F)

}

Upvotes: 2

TARehman
TARehman

Reputation: 6749

If you want to use basic R, you can write a simple function that does this and then use a for loop (or potentially apply) to do this.

func.widthmeans <- function(prefix,target.df) {

    active.df <- get(target.df)
    return(mean(active.df[grep(pattern=prefix,x=active.df$Name),"Width"]))
}

for(x in df1$Name) {

    df1[df1$Name==x,"MeanWidth"] <- func.widthmeans(prefix=x,target.df="df2")
}
rm(x)
df1

  Name Diam  Rad MeanWidth
1   t1   95 47.5  3600.000
2   t2  104 52.0  2366.667
3   t3   72 36.0  3500.000

This is a more generalized solution that is a function that accepts two data frames, one with the general names and one with the width data, like you described in the comments.

func.widthmeans <- function(data.df,width.df) {

    for(x in data.df$Name) {

        data.df[data.df$Name==x,"MeanWidth"] <- mean(width.df[grep(pattern=x,x=width.df$Name),"Width"])
    }
    rm(x)

    return(data.df)
}
func.widthmeans(data.df = df1,width.df = df2)

  Name Diam  Rad MeanWidth
1   t1   95 47.5  3600.000
2   t2  104 52.0  2366.667
3   t3   72 36.0  3500.000

Upvotes: 1

janos
janos

Reputation: 124734

Here's a solution to get df3 using sapply:

df3 <- df1
df3$WidthMean <- sapply(df1$Name, function(name) {
  pattern = paste('^', name, sep='')
  mean(subset(df2, grepl(pattern, Name))$Width)
})
df3

Outputs:

  Name Diam  Rad WidthMean
1   t1   95 47.5  3600.000
2   t2  104 52.0  2366.667
3   t3   72 36.0  3500.000

If you have 75 dfs that look like df1, you can use lapply to do them all:

lapply(list(df1, df1), function(df) {
  df_out <- df
  df_out$WidthMean <- sapply(df$Name, function(name) {
    pattern = paste('^', name, sep='')
    mean(subset(df2, grepl(pattern, Name))$Width)
  })
  df_out
})

Outputs:

[[1]]
  Name Diam  Rad WidthMean
1   t1   95 47.5  3600.000
2   t2  104 52.0  2366.667
3   t3   72 36.0  3500.000

[[2]]
  Name Diam  Rad WidthMean
1   t1   95 47.5  3600.000
2   t2  104 52.0  2366.667
3   t3   72 36.0  3500.000

Upvotes: 2

Related Questions