mergeVote: Merge Roll Call Vote

Description

Merge roll call vote record with a data.frame containing other information. The vote records are typically incomplete, so match first on houseSenate and surname. If this match is incomplete, try using givenName. If that fails, try state and district, which may not always be present in vote.

Usage

mergeVote(x, vote, Office="House", vote.x, check.x=TRUE)

Arguments

a data.frame whose columns include Office, surname, and givenName.

vote

a data.frame with column names which when forced tolower would match surname, givenname, and vote. However, the givenname may not be complete, so use it only if the surname is not sufficient.

Office

Either "House" or "Senate"; ignored if vote includes a column Office.

vote.x

name of a column of x containing a vote to be updated with the vote column of the vote data.frame. If missing and x has a column with a name matching "vote", then vote.x is that column. If missing but x has no such column, then append a column to x with the name of the vote column of the vote data.frame.

check.x

logical: If TRUE, check for rows of x[, vote.x] that are NOT in vote and throw an error if found.

Value

a data.frame with the same columns as x with its vote column modified per the vote argument.

Details

1. Parse vote.x to get the name of the column of x into which to write the vote column of the vote data.frame.

2. If the vote data.frame contains a column Office, ignore the Office argument. Otherwise, add the argument houseSenate as a column of vote.

3. Create keyx <- with(x, paste(Office, surname, sep=":")), keyx2 <- paste(keyx, givenName, sep=":"), keyx. <- paste(houseSenate, state, district, sep=":"), and similarly keyv, leyv2, and keyv. from vote.

4. Look for keyv in keyx. When a unique match is found, transfer the vote the vote column of x. When no match is found, try for keyv2 in keyx2 or keyv. in keyx. If those fail, print an error message with the information from vote on all failures and ask the user to add state and district informuation.

5. if(check.x), check for rows in x[, vote.x] that are NOT "notEligible" but are also not in vote: Throw an error if any are found.

Examples

Run this code

# NOT RUN {
##
## 1.  Test good cases
##
votetst <- data.frame(
  surName=c('Smith', 'Jones', 'Graves', 'Jsn', 'Jsn', 'Gay'),
  givenName=c("Sam", "", "", "John", "John", ''),
  votex=factor(c('Y', 'N', 'abstain', 'Y', 'Y', 'Y')),
  State=factor(rep(c("CA", "", "SC", "NY"), c(1, 2, 1, 2))),
  district=rep(c("13", "1", "2", "1"), c(1, 2, 2, 1)),
  stringsAsFactors=FALSE )

x1 <- data.frame(
  Office=factor(rep(c("House", "Senate"), e=8)),
  state=rep(c("NY", "SC", "SD", "CA", "AK", "AR", "NY", "NJ"), 2),
  District=rep(c("2", "2", "At Large", "13", "1", "9", "1", "3"), 2),
  surname=rep(c('Jsn', 'Jsn', 'Smith', 'Smith', 'Jones',
       'Graves', 'Rx', 'Agnew'), 2),
  givenName=rep(c("John D.", "John J.",
    "Samual", "Samual", "Mary", "Mary", "Susan", 'Spiro'), 2),
  don=1:16, stringsAsFactors=FALSE)

x1. <- mergeVote(x1, votetst)

x2 <- cbind(x1, votex=factor( rep(
   c('Y', 'notEligible', 'Y', 'N', 'abstain', 'Y', 'notEligible'),
   c(2,1,1,1,1,1,9) ) ) )

# }
# NOT RUN {
all.equal(x1., x2)
# }
# NOT RUN {
##
## 2.  Test a case with a vote error in x
##

x1a <- cbind(x1, voterr=rep(
     c('notEligible', 'Y', 'notEligible'), c(7, 1, 8)))

x1a. <- try(mergeVote(x1a, votetst))

# }
# NOT RUN {
class(x1a.)=='try-error'
# }
# NOT RUN {
# }