R/prep-license.R

In southwick-associates/salicprep: Prepare Agency Data

# preparing license data

#' Convert date variables from character to date format
#' 
#' A convenience function that also includes a summary of any input values
#' thate weren't converted to dates (i.e., missing in output).
#' 
#' @param df data frame that contains date variable
#' @param date_var date variable stored as character
#' @param fun function used to convert from character to date format
#' @family functions for preparing license data
#' @export
#' @examples 
#' # recode_date(cust, "dob", function(x) lubridate::ymd(str_sub(x, end = 10)))
recode_date <- function(df, date_var, fun) {
    if (!is.character(df[[date_var]])) {
        stop("The date_var must be a character type", call. = FALSE)
    }
    df$date_in <- df[[date_var]]
    df[[date_var]] <- fun(df[[date_var]])
    
    # print a check of values that failed to code as date
    cat("Counts of 'failed to parse' by input date value:\n")
    if (any(is.na(df[[date_var]]))) {
        df %>%
            filter(is.na(.data[[date_var]])) %>% 
            count(.data$date_in) %>% 
            arrange(desc(.data$n)) %>%
            data.frame() %>% 
            print() 
    }
    select(df, -.data$date_in)
}

#' Convert all date columns in a data frame to character
#' 
#' A convenience function to prepare for sqlite, which doesn't translate R dates 
#' when writing data. It's a very slow operation, so not efficient with regard to
#' computation, but the alternative is to specify as character on load; 
#' a tedious thing to do with a larger number of variables.
#' 
#' @param df input data frame
#' @family functions for preparing license data
#' @export
#' @examples 
#' # sale <- date_to_char(sale)
date_to_char <- function(df) {
    is_date_col <- sapply(names(df), function(x) {
        lubridate::is.POSIXt(df[[x]]) || lubridate::is.POSIXlt(df[[x]]) || 
            lubridate::is.POSIXct(df[[x]]) || lubridate::is.Date(df[[x]]) 
    }) %>% 
        unlist()
    date_cols <- names(df)[is_date_col]
    df[date_cols] <- lapply(df[date_cols], as.character)
    df
}

#' Compare count of lines in raw file to imported data frame
#' 
#' Generally speaking the raw file will have one more line (header) than the 
#' data frame has rows. This function is mostly a wrapper for
#' \code{\link[salic]{count_lines_textfile}}
#' 
#' @param df data frame into which raw_file was loaded
#' @param raw_file path to raw data file
#' @family functions for preparing license data
#' @export
#' @examples 
#' # check_raw_lines(cust, "file/path/to/raw/cust/text/file")
check_raw_lines <- function(df, raw_file) {
    raw_count <- salic::count_lines_textfile(raw_file)
    df_count <- nrow(df)
    data.frame(
        name = c("raw_file", "df", "difference"),
        row_count = c(raw_count, df_count, raw_count - df_count),
        stringsAsFactors = FALSE
    )
}