In R, how can I copy rows from one dataframe to another when the df being copied to has 2 additional columns?

Question

I have a tab delimited text file with 12 columns that I am uploading to my program. I go on to create another dataframe with a structure similar to the one uploaded and add 2 more columns to it.

excelfile = read.delim(ExcelPath)
matchedPictures<- excelfile[0,]
matchedPictures$beforeName <- character()
matchedPictures$afterName <- character()

Now I have a function in which I do the following:

Based on a condition, I obtain the row number pictureMatchNum of the row I need to copy from excelfile to matchedPictures.

I should then copy the row from excelfile to matchedPictures. I tried a couple of different ways so far.

a.

rowNumber = nrow(matchedPictures) + 1
matchedPictures[rowNumber,1:12] <<- excelfile[pictureMatchNum,1:12]

b.

matchedPictures[rowNumber,1:12] <<- rbind(matchedPictures, excelfile[pictureWordMatches,1:12], make.row.names = FALSE)

2a. doesn't seem to work because it copies the indices from the excelfileand uses them as row names in the matchedPictures - which is why I decided to go with rbind

2b. doesn't seem to work because rbind needs to have the columns be identical and matchedPictureshas 2 extra columns.

EDIT START - Including reproducible example.

Here is some reproducible code (with fewer columns and fake data)

excelfile <- data.frame(x = letters, y = words[length(letters)], z= fruit[length(letters)] )
matchedPictures <- excelfile[0,]
matchedPictures$beforeName <- character()
matchedPictures$afterName <- character()

pictureMatchNum1 = match(1, str_detect("A", regex(excelfile$x, ignore_case = TRUE)))
rowNumber1 = nrow(matchedPictures) + 1

pictureMatchNum2 = match(1, str_detect("D", regex(excelfile$x, ignore_case = TRUE)))
rowNumber2 = nrow(matchedPictures) + 1

The 2 options I tried are

2a.

matchedPictures[rowNumber1,1:3] <<- excelfile[pictureMatchNum1,1:3]
matchedPictures[rowNumber1,"beforeName"] <<- "xxx"
matchedPictures[rowNumber1,"afterName"] <<- "yyy"

matchedPictures[rowNumber2,1:3] <<- excelfile[pictureMatchNum2,1:3]
matchedPictures[rowNumber2,"beforeName"] <<- "uuu"
matchedPictures[rowNumber2,"afterName"] <<- "www"

OR

2b.

matchedPictures[rowNumber1,1:3] <<- rbind(matchedPictures, excelfile[pictureMatchNum1,1:3], make.row.names = FALSE)
matchedPictures[rowNumber1,"beforeName"] <<- "xxx"
matchedPictures[rowNumber1,"afterName"] <<- "yyy"

matchedPictures[rowNumber2,1:3] <<- rbind(matchedPictures, excelfile[pictureMatchNum2,1:3], make.row.names = FALSE)
matchedPictures[rowNumber2,"beforeName"] <<- "uuu"
matchedPictures[rowNumber2,"afterName"] <<- "www"

EDIT END

Additionally, I have also seen the suggestions in many places that rather than using empty dataframes, one should have vectors and append data to the vectors and then combine them into a dataframe. Is this suggestion valid when I have so many columns and would need to have 14 separate vectors and copy each one of them individually?

What can I do to make this work?

Please provide a reproducible example to illustrate your problem. http://stackoverflow.com/questions/5963269/how-to-make-a-great-r-reproducible-example — Adam Quek, Apr 13 '17 at 10:23
@AdamQuek Thank you for your suggestion to make my question better. I have made the suggested improvement. — ag14, Apr 13 '17 at 10:58

score 0 · Answer 1 · answered Apr 13 '17 at 11:59

You could

first determine the row indices of excelfile that match your criteria
extract these rows
then generate the data to fill your columns beforeName and afterName
then append these columns to your new data frame

Example:

excelfile <- data.frame(x = letters, y = words[length(letters)], 
    z = fruit[length(letters)])
    ## Vector of patterns:
patternVec <- c("A", "D", "M")
## Look for appropriate rows in file 'excelfile':
indexVec <- vapply(patternVec, 
        function(myPattern) which(str_detect(myPattern, 
                    regex(excelfile$x, ignore_case = TRUE))), integer(1))
## Extract these rows:
matchedPictures <- excelfile[indexVec,]
## Somehow generate the data for columns 'beforeName' and 'afterName':
## I do not know how this information is generated so I just insert 
## some dummy code here:
beforeNameVec <- c("xxx", "uuu", "mmm")
afterNameVec <- c("yyy", "www", "nnn")
## Then assign these variables:
matchedPictures$beforeName <- beforeNameVec
matchedPictures$afterName <- afterNameVec

matchedPictures
# x   y           z beforeName afterName
# a air dragonfruit        xxx       yyy
# d air dragonfruit        uuu       www
# m air dragonfruit        mmm       nnn

score 0 · Answer 2 · answered Apr 13 '17 at 11:59

You can make this much simpler by using dplyr

library(dplyr)
library(stringr)

excelfile <- data.frame(x = letters, y = words[length(letters)], z= fruit[length(letters)],
stringsAsFactors = FALSE ) #add stringsAsFactors to have character columns

pictureMatch <- excelfile %>%
  #create a match column
  mutate(match = ifelse(str_detect(x,"a") | str_detect(x,'d'),1,0)) %>% 
  #filter to only the columns that match your condition
  filter(match ==1)

pictureMatch <- pictureMatch[['x']] #convert to a vector

matchedPictures <- excelfile %>%
  filter(x %in% pictureMatch) %>% #grab the rows that match your condition
  mutate(beforeName = c('xxx','uuu'), #add your names
     afterName = c('yyy','www'))

In R, how can I copy rows from one dataframe to another when the df being copied to has 2 additional columns?

2 Answers2