Support strings_as_factors option

[eddelbuettel]

This PR adds an optional binary toggle strings_as_factors along with getters and setters to permit conversion of character columns to factors.

The default value is the value of the (existing base R) option("stringsAsFactors"). A quick demonstration on the 'penguins' dataset (with a added index column) shows how the three character column are converted which makes for example summary() for informative for counts:

edd@rob:~/git/tiledb-r(de/sc-14685/strings_as_factors)$ r -ltiledb -e'arr <- tiledb_array("/tmp/tiledb/penguinsNew", strings_as_factors=FALSE); res <- arr[]; print(summary(res))'
 __tiledb_rows     species             island          bill_length_mm bill_depth_mm  flipper_length_mm  body_mass_g       sex                 year     
 Min.   :  1.0   Length:344         Length:344         Min.   :32.1   Min.   :13.1   Min.   :172       Min.   :2700   Length:344         Min.   :2007  
 1st Qu.: 86.8   Class :character   Class :character   1st Qu.:39.2   1st Qu.:15.6   1st Qu.:190       1st Qu.:3550   Class :character   1st Qu.:2007  
 Median :172.5   Mode  :character   Mode  :character   Median :44.5   Median :17.3   Median :197       Median :4050   Mode  :character   Median :2008  
 Mean   :172.5                                         Mean   :43.9   Mean   :17.2   Mean   :201       Mean   :4202                      Mean   :2008  
 3rd Qu.:258.2                                         3rd Qu.:48.5   3rd Qu.:18.7   3rd Qu.:213       3rd Qu.:4750                      3rd Qu.:2009  
 Max.   :344.0                                         Max.   :59.6   Max.   :21.5   Max.   :231       Max.   :6300                      Max.   :2009  
                                                       NA's   :2      NA's   :2      NA's   :2         NA's   :2                                       
edd@rob:~/git/tiledb-r(master)$ r -ltiledb -e'arr <- tiledb_array("/tmp/tiledb/penguinsNew", strings_as_factors=TRUE); res <- arr[]; print(summary(res))'
 __tiledb_rows        species          island    bill_length_mm bill_depth_mm  flipper_length_mm  body_mass_g       sex           year     
 Min.   :  1.0   Adelie   :152   Biscoe   :168   Min.   :32.1   Min.   :13.1   Min.   :172       Min.   :2700   female:165   Min.   :2007  
 1st Qu.: 86.8   Chinstrap: 68   Dream    :124   1st Qu.:39.2   1st Qu.:15.6   1st Qu.:190       1st Qu.:3550   male  :168   1st Qu.:2007  
 Median :172.5   Gentoo   :124   Torgersen: 52   Median :44.5   Median :17.3   Median :197       Median :4050   NA's  : 11   Median :2008  
 Mean   :172.5                                   Mean   :43.9   Mean   :17.2   Mean   :201       Mean   :4202                Mean   :2008  
 3rd Qu.:258.2                                   3rd Qu.:48.5   3rd Qu.:18.7   3rd Qu.:213       3rd Qu.:4750                3rd Qu.:2009  
 Max.   :344.0                                   Max.   :59.6   Max.   :21.5   Max.   :231       Max.   :6300                Max.   :2009  
                                                 NA's   :2      NA's   :2      NA's   :2         NA's   :2                                 
edd@rob:~/git/tiledb-r(master)$ 

[shortcut-integration]

This pull request has been linked to Shortcut Story #14685: Support strings_as_factors for character columns.

[aaronwolen]

Very nice addition!

[aaronwolen]

Smart to use the global option here!

[eddelbuettel]

Thanks 😀 It's a not-uncommon pattern I picked up 'somewhere' along the way; I do recall at some point a few years back Duncan Temple-Lang helped with a snippet like this. The fallback in getOption() is also golden.