CODEWITHSUNDEEP
rparsing
wes
wes

Reputation: 113

Parsing out data from one variable using complex rules in R

I am importing data into R from another source (i.e., I cannot easily change the in-coming format/values).

Among the variables is one that include one or more of these possible values:

all within the same "cell" so that possible data look like:

Sample Input Data Frame (df)

df <- read.table(text =
"row lives.with.whom
  1  'Mother (biological mother, foster mother, step mother, etc.), Father (biological father, foster father, step father, etc.), Grandparent(s) (biological, foster, step, etc.), Brother(s) older than 18, Sister(s) older than 18, Other adults (aunts, uncles, etc.)'
  2  ''
  3  'Mother (biological mother, foster mother, step mother, etc.), Sister(s) older than 18'
  4  'Mother (biological mother, foster mother, step mother, etc.), Father (biological father, foster father, step father, etc.)'", header = T)

Within R, how could I efficiently create rules to parse out these responses into separate columns, one column for each type of family member, so that the output would look like this:

Sample Output Data Frame

mother <- c(1,0,1,1)
father <- c(1,0,0,1)
adult.brother <- c(1,0,0,0)
adult.sister <- c(1,0,1,0)
grandparent <- c(1,0,0,0)
other.adult <- c(1,0,0,0)
output.df <- cbind(mother, father, adult.brother, adult.sister, grandparent, other.adult)
colnames(output.df) <- c("Mother", "Father", "Brother", "Sister", "Grandparent", "Other adult")
output.df

     Mother Father Brother Sister Grandparent Other adult
[1,]      1      1       1      1           1           1
[2,]      0      0       0      0           0           0
[3,]      1      0       0      1           0           0
[4,]      1      1       0      0           0           0

Upvotes: 0

Views: 76

Answers (3)

Suhas Hegde
Suhas Hegde

Reputation: 416

I made some assumptions and tried to solve it.

library(tidyr)
library(dplyr)
# create nested lists with names of mothers and fathers for two ppl
mother <- list(list("bio_1","step_1","foster_1"), list("bio_2", "stp_2", "foster_2"))
father <- list(list("bio_1", "foster_1", "other_1"), list("bio_2", "stp_2", "foster_2"))

# convert to data frame
test_object <- data_frame(person = c(1,2),mother,father)

# print 
test_object

# A tibble: 2 x 3
  person mother     father    
   <dbl> <list>     <list>    
1      1 <list [3]> <list [3]>
2      2 <list [3]> <list [3]>

# first unnest the lists and get to the inner list
# then convert from wide to long form data
# do another unnnest to get the actual data in the long format
test_object %>%
  unnest(.) %>%
    gather(data = ., key = relationship, value = name, -person) %>%
      unnest() -> test_object
    
    test_object
# A tibble: 12 x 3
   person relationship name    
    <dbl> <chr>        <chr>   
 1      1 mother       bio_1   
 2      1 mother       step_1  
 3      1 mother       foster_1
 4      2 mother       bio_2   
 5      2 mother       stp_2   
 6      2 mother       foster_2
 7      1 father       bio_1   
 8      1 father       foster_1
 9      1 father       other_1 
10      2 father       bio_2   
11      2 father       stp_2   
12      2 father       foster_2

Here are links to tidyverse and data.table that contain a lot packages and functions to solve most of your data-carpentry/wrangling issues.

Upvotes: 1

Ankur
Ankur

Reputation: 151

Try this:

rel<-list("Mother", "Father", "Brother", "Sister", "Grandparent", "Other adult")

for(i in 1:6){
  df$i<-if_else(grepl(rel[[i]],df$lives.with.whom),1,0)
  colnames(df)[i+2]<-rel[[i]]
}

Upvotes: 1

Maurits Evers
Maurits Evers

Reputation: 50668

Here is a tidyverse option that should get you started

library(tidyverse)
rel <- list("Mother", "Father", "Brother", "Sister", "Grandparent", "Other adult")
names(rel) <- unlist(rel)
bind_cols(df[, 1, drop = F], map(rel, ~+str_detect(tolower(df[, 2]), tolower(.x))))
#  row Mother Father Brother Sister Grandparent Other adult
#1   1      1      1       1      1           1           1
#2   2      0      0       0      0           0           0
#3   3      1      0       0      1           0           0
#4   4      1      1       0      0           0           0

Sample data

df <- read.table(text =
    "row lives.with.whom
  1  'Mother (biological mother, foster mother, step mother, etc.), Father (biological father, foster father, step father, etc.), Grandparent(s) (biological, foster, step, etc.), Brother(s) older than 18, Sister(s) older than 18, Other adults (aunts, uncles, etc.)'
  2  ''
  3  'Mother (biological mother, foster mother, step mother, etc.), Sister(s) older than 18'
  4  'Mother (biological mother, foster mother, step mother, etc.), Father (biological father, foster father, step father, etc.)'", header = T)

Upvotes: 1

Related Questions