collapse: Collapse objects
In setempler/miscset: Miscellaneous Tools Set

Description Usage Arguments Details Author(s) Examples

Collapse objects as in the paste function.

collapse(x, sep, ...)

## Default S3 method:
collapse(x, sep = "", ..., .unique = FALSE,
  .sort = FALSE, .decreasing = FALSE)

## S3 method for class 'data.frame'
collapse(x, sep = "", by = names(x), ...,
  .unique = FALSE, .sort = FALSE, .decreasing = FALSE, .unlist = FALSE,
  .sortby = FALSE)

`x`	Any R object.
`sep`	A character string to separate value columns. `NULL` retains a vector.
`...`	Forwarded to or from other methods.
`.unique`	Logical, return only unique values.
`.sort`	Logical, sort the values.
`.decreasing`	Logical, if sorting, then by decreasing values.
`by`	Column names to split data frame by, before applying collapse on each remaining column within each piece. Using the default (all columns), then `unique(x)` is returned. Columns can be specified by names or integer with the column numbers. Using `0` or `NULL` collapses all columns.
`.unlist`	Logical, if value columns need to be unlisted before collapsing.
`.sortby`	Logical, sort the output on the `by` columns. This applies , If `x` was a `data.table`, then the keys are set as the `by` values.

For the data.frame method, x is converted to a data.table before applying the piece- and column-wise collapses. If the input is already inheriting from data.table, then the class is retained.
.sortby is causing setkeyv(x, by) to be applied to x after converting to a data.table.

Sven E. Templer

#

### some data

set.seed(12)
s <- s2 <- sample(LETTERS[1:4], 9, replace = TRUE)
s2[1:2] <- rev(s2[1:2])
d <- data.frame(group = rep(letters[c(3,1,2)], each = 3), 
                value = s,
                level = factor(s2),
                stringsAsFactors = FALSE)

### collapse vectors

collapse(letters)
collapse(1:3)               # coerced to character
collapse(LETTERS[1:5], "-") # separated by '-'

### collapse data.frames

# by all columns (same as unique)
collapse(d)
# by a grouping column
collapse(d, by = 1)
# by multiple, but not all columns
collapse(d, by = c("group", "value"))
# return single row
collapse(d, by = 0)
# return single row, unique and sorted values
collapse(d, by = 0, .unique = TRUE, .sort = TRUE)

#