Common interface: read_(f)ods, write_(f)ods, list_(f)ods_sheets

[chainsawriot]

As the only difference between read_ods and read_fods (as well as other functions such as list_ods_sheets and the upcoming write_ods) is flat when calling .read_ods, I was thinking about making read_ods and co. work like readxl::read_excel().

We should have a function like excel_format() https://github.com/tidyverse/readxl/blob/main/R/excel-format.R

For our case, it is mostly file extension: ods vs fods/xml; and then file signature (zip file's as.raw(c( "0x50", "0x4B", "0x03", "0x04" ))).

A rough implementation (unlike readxl, I don't want it to be exported):

.determine_ods_format <- function(path, guess = FALSE) {
    ext <- tolower(tools::file_ext(path))
    formats <- c(
        ods = "ods",
        fods = "fods",
        xml = "fods"
    )
    if (!isTRUE(guess)) {
        return(unname(formats[ext]))
    }
    zip_sig <- as.raw(c(
        "0x50", "0x4B", "0x03", "0x04"
    ))
    if (identical(zip_sig, readBin(path, n = 4, what = "raw"))) {
        return("ods")
    } else {
        return("fods")
    }
}

.determine_ods_format(readODS::write_ods(iris))
#> [1] "ods"
## will be available soon
##.determine_ods_format(readODS::write_fods(iris)) # fods

.determine_ods_format(readODS::write_ods(iris, tempfile(fileext = ".fods"))) ## bitte nicht nachmachen.
#> [1] "fods"
.determine_ods_format(readODS::write_ods(iris, tempfile(fileext = ".fods")), guess = TRUE)
#> [1] "ods"

^{Created on 2023-08-21 with reprex v2.0.2}

read_ods, for example, would then be:

read_ods <- function(path,
                     sheet = 1,
                     col_names = TRUE,
                     col_types = NULL,
                     na = "",
                     skip = 0,
                     formula_as_formula = FALSE,
                     range = NULL,
                     row_names = FALSE,
                     strings_as_factors = FALSE,
                     verbose = FALSE,
                     as_tibble = TRUE,
                     .name_repair = "unique",
                     guess = FALSE) {
    ## Should use match.call but there's a weird bug if one of the variable names is 'file'
    .read_ods(path = path,
        sheet = sheet,
        col_names = col_names,
        col_types = col_types,
        na = na,
        skip = skip,
        formula_as_formula = formula_as_formula,
        range = range,
        row_names = row_names,
        strings_as_factors = strings_as_factors,
        verbose = verbose,
        as_tibble = as_tibble,
        .name_repair = .name_repair,
        flat = .determine_ods_format(path, guess = guess) == "fods")
}

[chainsawriot]

Or:

read_ods <- function(path,
                     sheet = 1,
                     col_names = TRUE,
                     col_types = NULL,
                     na = "",
                     skip = 0,
                     formula_as_formula = FALSE,
                     range = NULL,
                     row_names = FALSE,
                     strings_as_factors = FALSE,
                     verbose = FALSE,
                     as_tibble = TRUE,
                     .name_repair = "unique",
                     ods_format = "auto"
                     guess = FALSE) {
    if (ods_format == "auto") {
         ods_format <- .determine_ods_format(path, guess = guess)
    }
    ## Should use match.call but there's a weird bug if one of the variable names is 'file'
    .read_ods(path = path,
        sheet = sheet,
        col_names = col_names,
        col_types = col_types,
        na = na,
        skip = skip,
        formula_as_formula = formula_as_formula,
        range = range,
        row_names = row_names,
        strings_as_factors = strings_as_factors,
        verbose = verbose,
        as_tibble = as_tibble,
        .name_repair = .name_repair,
        flat = ods_format == "fods")
}

so that one can do crazy thing like:

read_ods("it_is_actually_fods_but_I_dont_want_you_to_guess.ods", ods_format = "fods")