make_dull: Hide dull columns in data frames

In mvbutils: General utilities, workspace organization, code and doc editing, live package maintenance, etc

Hide dull columns in data frames

Description

make_dull AKA make.dull adds a "dull" S3 class to designated columns in a data.frame. When the data.frame is printed, entries in those columns will show up just as "...". Useful for hiding long boring stuff like matrices with loads of columns, nucleotide sequences, MD5 sums, and filenames. Columns will still print clearly and behave properly if manually extracted. You can remove dullness via undull; see Examples.

The dull class has methods for format (used when printing a data.frame) and [, so that dullness is perpetuated.

Usage

make_dull(df, cols)

Arguments

df	a data.frame
cols	columns to designate

Details

Ask yourself: do you really want details of a function called make_dull? Life may be sweet but it is also short.

Actually, I've had to add something "for the record", but you probably don't want to read this. Just prepending "dull" to the "class" attribute of a column (see oldClass) can mask normal S3 dispatch— matrices being a case in point. Therefore, make_dull now also adds (part of) the implicit class of the column after "dull". In quasi-English, what that means is that a matrix will acquire explicit class attribute c( "dull", "matrix", "array"), whereas a normal non-dull matrix would have a NULL "class" attribute but an implicit class of c("matrix","array"). This means you can still do e.g. isSymmetric( x$dullmat) and inherits( x$dullmat, "matrix"). Other semi-exotic objects (including array, which AFAICR only semi-works inside dataframes, list which I think works OK, and user-defined classes) get similar treatment.

This Sort Of Thing is why it's marginally worth having an explicit undull function, as opposed to just unclass or eg oldClass(excol) <- oldClass( excol) %except% "dull". The former would destroy a user-defined class; the latter would leave superfluous matrix and array elements in an unnecessary "class" attribute (although I'm not sure that causes any practical problems).

Sigh... we are deep in the S3wamplands here; "the perfect is the enemy of the good" where S3 is concerned, so we just gotta put up with some of the murky stuff. That's OK. Be careful what you wish for: ie S4!

More details

make_dull is both autologous and idempotent.

Value

A modified data.frame

Examples

# Becos more logical syntax:
rsample <- function (n = length(pop), pop, replace = FALSE, prob = NULL){
  pop[sample(seq_along(pop) - 1, size = n, replace = replace, prob = prob) + 1]
}
df <- data.frame( x=1:3,
    y=apply( matrix( rsample( 150, as.raw( 33:127), rep=TRUE), 50, 3), 2, rawToChar),
    stringsAsFactors=FALSE) # s.A.F. value shouldn't matter
df # zzzzzzzzzzzzzzz
df <- make_dull( df, 'y')
df # wow, exciting!
df$y # zzzzzzzzzzzzzz
undull( df$y) # no class attrib now
df$ZZZ <- matrix( 1:99, nrow=3, ncol=33)
df # boooring
class( df$ZZZ)
oldClass( df$ZZZ)
df <- make_dull( df, 'ZZZ')
df # whew
class( df$ZZZ)
oldClass( df$ZZZ)
ZZZ <- df$ZZZ
# Suddenly it is interesting! So...
ZZZ <- undull( ZZZ)
class( ZZZ)
oldClass( ZZZ)

mvbutils documentation built on May 25, 2026, 5:09 p.m.