CODEWITHSUNDEEP
r
user124123
user124123

Reputation: 1683

create a multiple data frames from a single data frame in r

I've got a large data frame finaldata and i would like to produce a bunch of other smaller data frames explanatory1, explanatory2 e.t.c.... consisting of 10 columns each from finaldata

I'm trying to do this using a for loop but its throwing me an error attempt to apply non function

for(i in 1:length(finaldata)/10) {
  nam <- paste("explanatory", i, sep = "")
  assign(nam, finaldata[,10(i):10(i)+10])
}

I have also tried

for(i in 1:length(finaldata)/10){
  assign(paste("explanatory",i,sep=""),finaldata[,10(i):10(i)+10])}

But this gave me the same error, from what I understand the error is being caused by my passing finaldata[,10(i):10(i)+10] as an argument to assign, but I don't see why it wouldn't work ina for loop, or be any different from passing finaldata[,10:10+10]

Any help would be greatly appreciated!

Upvotes: 0

Views: 657

Answers (4)

user124123
user124123

Reputation: 1683

This did exactly it, thanks to @canary_in_the_data_mine. Choose "number_of_dataframes" to be some factor of "finaldata" that you want for your purpose, then:

number_of_dataframes <- ceiling(ncol(finaldata) / 5)
current_column <- 1
for (i in 1:number_of_dataframes) {
  start_column <- current_column
  end_column <- current_column + 5
  assign(paste0(explanatory,i), finaldata[,start_column:end_column])
  current_column <- end_column + 1
}

The only change I made was to end_column.

Upvotes: 1

vrajs5
vrajs5

Reputation: 4126

You were almost there... Try this... 

for(i in 1:ncol(finaldata)/10) {
  nam <- paste0("explanatory")
  if((10*(i - 1)+10) > ncol(finaldata)){
        assign(nam, finaldata[,(10*(i-1) +1):ncol(finaldata)])
  }else{
        assign(nam, finaldata[,(10*(i-1) +1):(10*(i - 1)+10)])
  }
}

Upvotes: 1

canary_in_the_data_mine
canary_in_the_data_mine

Reputation: 2393

Create sample data to play with: 

df <- data.frame(matrix(vector(), 10, 33))

Find the number of dataframes you're going to create: 

number_of_dataframes <- ceiling(ncol(df) / 10)

Loop through the dataframes, finding a range of columns to use for creating that individual dataframe. Use assign to give each one a unique name: 

current_column <- 1
for (i in 1:number_of_dataframes) {
  start_column <- current_column
  end_column <- min(current_column + 9, ncol(df))
  assign(paste0("df",i), df[ , start_column:end_column])
  current_column <- end_column + 1
}

The min check makes sure you don't attempt to assign more columns than existed in the original dataframe.

Upvotes: 2

agstudy
agstudy

Reputation: 121568

Using split:

ll <- lapply(split(colnames(finaldata),rep(seq_len(ncol(finaldata)/10),each=10)),
       function(x)finaldata[,x])

This will create a list. But You can extract from it separate variables (not recommanded) :

ll <- setNames(ll,paste0("explanatory",seq_along(ll)))
list2env(ll)

Upvotes: 4

Related Questions