CODEWITHSUNDEEP
rsurvey
EML
EML

Reputation: 671

Loop though columns and print tables using survey weights from survey package

I am trying to loop though columns of a dataset and create proportion tables. The survey package requires a '~' before variables making it tricky to loop through columns.

Data and design:

library(survey)
dat <- data.frame(id=c(1,2,3), weight=c(0,2,0.1), var1=c(2,3,5), var2=c(2,6,7))
design <- svydesign(id=~1, weights=~weight, data=dat)

Attempt: 

cols <- c("var1", "var2") # columns to loop through
for(i in seq_along(cols)) {
prop.table(svytable(~i, design))
}

I get the following error message:

Error in model.frame.default(formula = weights ~ i, data = model.frame(design)) : variable lengths differ (found for 'i')

Thank you

Upvotes: 2

Views: 523

Answers (1)

deepseefan
deepseefan

Reputation: 3791

Acknowledging the use of bquote as a partial substitution in expression from Ben's comment above, you could modify your for loop as follows:

cols <- c("var1", "var2") # columns to loop through
for(i in  seq_along(cols)){
  print(prop.table(svytable(bquote(~.(as.name(cols[i]))), design)))
} 
# var1
#          2          3          5 
# 0.00000000 0.95238095 0.04761905 
# var2
#          2          6          7 
# 0.00000000 0.95238095 0.04761905

Data

library(survey)
dat <- data.frame(id=c(1,2,3), weight=c(0,2,0.1), var1=c(2,3,5), var2=c(2,6,7))
design <- svydesign(id=~1, weights=~weight, data=dat)

Upvotes: 2

Related Questions