calculating number of days between 2 columns of dates in data frame

Without your seeing your data (you can use the output of dput(head(survey)) to show us) this is a shot in the dark:

survey <- data.frame(date=c("2012/07/26","2012/07/25"),tx_start=c("2012/01/01","2012/01/01"))

survey$date_diff <- as.Date(as.character(survey$date), format="%Y/%m/%d")-
                  as.Date(as.character(survey$tx_start), format="%Y/%m/%d")
survey
       date   tx_start date_diff
1 2012/07/26 2012/01/01  207 days
2 2012/07/25 2012/01/01  206 days

You could find the difference between dates in columns in a data frame by using the function difftime as follows:

df$diff_in_days<- difftime(df$datevar1 ,df$datevar2 , units = c("days"))

You need to use the as.Date formats correctly.

Eg.

x = '2012/07/25'
xd = as.Date(x,'%Y/%m/%d')
xd    # Prints "2012-07-25"

R date formats are similary to *nix ones.

Doing a typeof(xd) shows it as a double ie. days since 1970.

In Ronald's example, if the date formats are different (as displayed below) then modify the format parameter

survey <- data.frame(date=c("2012-07-26","2012-07-25"),tx_start=c("2012-01-01","2012-01-01"))

survey$date_diff <- as.Date(as.character(survey$date), format="%Y-%m-%d")-
              as.Date(as.character(survey$tx_start), format="%Y-%m-%d")

survey:

   date      tx_start     date_diff
1 2012-07-26     2012-01-01    207 days
2 2012-07-25     2012-01-01    206 days