cols argument for unique.data.table

[MichaelChirico]

Closes #5243.

Extending the benchmark in the issue to include the new approach:

NN = 2e7
DT = data.table(grp = sample(c(letters, LETTERS, 0:9), NN, TRUE))
JJ = 100
for (jj in seq_len(JJ)) set(DT, NULL, paste0("V", jj), rnorm(NN))

BY_COLS = "grp"
KEEP_COLS = paste0("V", 1:5)

f1 <- function() DT[, unique(.SD, by = BY_COLS), .SDcols = c(BY_COLS, KEEP_COLS)]
f2 <- function() unique(DT[, .SD, .SDcols = c(BY_COLS, KEEP_COLS)], by = BY_COLS)
f3 <- function() unique(DT, by = BY_COLS)[, .SD, .SDcols = c(BY_COLS, KEEP_COLS)]
f4 <- function() DT[, head(.SD, 1L), by = BY_COLS, .SDcols = KEEP_COLS]
f5 <- function() unique(DT, by=BY_COLS, cols=KEEP_COLS)

bench::mark(min_iterations = 10L, f1(), f2(), f3(), f4(), f5())
# A tibble: 5 x 13
#  expression     min  median `itr/sec` mem_alloc `gc/sec` n_itr  n_gc total_time
#   <bch:expr> <bch:t> <bch:t>     <dbl> <bch:byt>    <dbl> <int> <dbl>   <bch:tm>
# 1 f1()        75.7ms  77.7ms     12.9         NA    0        10     0   776.75ms
# 2 f2()       443.2ms 445.5ms      2.24        NA    0.960     7     3      3.12s
# 3 f3()        75.5ms  75.8ms     13.1         NA    0        10     0   763.13ms
# 4 f4()         191ms 219.7ms      4.41        NA    0.490     9     1      2.04s
# 5 f5()        75.1ms  76.5ms     13.0         NA    0        10     0   768.13ms
# … with 4 more variables: result <list>, memory <list>, time <list>, gc <list>

So, we've achieved the goal of having a clean API for doing something fast previously available only through more convoluted code.

[codecov]

Codecov Report

Merging #5244 (255ea43) into master (0404ed8) will increase coverage by 0.00%.
The diff coverage is 100.00%.

@@           Coverage Diff           @@
##           master    #5244   +/-   ##
=======================================
  Coverage   99.50%   99.50%           
=======================================
  Files          77       77           
  Lines       14645    14647    +2     
=======================================
+ Hits        14573    14575    +2     
  Misses         72       72           

Impacted Files	Coverage Δ
R/duplicated.R	100.00% <100.00%> (ø)

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 0404ed8...255ea43. Read the comment docs.

[ben-schwen]

If we include it in unique.data.table shouldn't we also include it in duplicated.data.table and anyDuplicated.data.table since these go hand in hand?

[MichaelChirico]

not really -- cols only affects the output object. for both of those functions, the output wouldn't be affected, right?