rbindlist gives wrong results when it is applied to difftime objects with different units.

[lgong-rms]

# Minimal reproducible example

library(data.table)

difftime_diffunits <- list(
  structure(list(
    start_time = "2020-06-02 05:12:14.652", end_time = "2020-06-02 06:01:24.753", 
    diff_time = structure(49.1683500011762, class = "difftime", units = "mins")), 
    row.names = c(NA, -1L), class = c("data.table", "data.frame")), 
  structure(list(
    start_time = "2020-06-02 05:12:14.656", end_time = "2020-06-02 06:02:25.045", 
    diff_time = structure(50.1731500029564, class = "difftime", units = "mins")), 
    row.names = c(NA, -1L), class = c("data.table", "data.frame")), 
  structure(list(
    start_time = "2020-06-02 05:12:14.973", end_time = "2020-06-02 06:32:28.081", 
    diff_time = structure(1.33697444445557, class = "difftime", units = "hours")), 
    row.names = c(NA, -1L), class = c("data.table", "data.frame")), 
  structure(list(
    start_time = "2020-06-02 05:12:14.973", end_time = "2020-06-02 06:24:51.367", 
    diff_time = structure(1.21010944445928, class = "difftime", units = "hours")), 
    row.names = c(NA, -1L), class = c("data.table", "data.frame")))

# [[1]]
# start_time                end_time     diff_time
# 1: 2020-06-02 05:12:14.652 2020-06-02 06:01:24.753 49.16835 mins
# 
# [[2]]
# start_time                end_time     diff_time
# 1: 2020-06-02 05:12:14.656 2020-06-02 06:02:25.045 50.17315 mins
# 
# [[3]]
# start_time                end_time      diff_time
# 1: 2020-06-02 05:12:14.973 2020-06-02 06:32:28.081 1.336974 hours
# 
# [[4]]
# start_time                end_time      diff_time
# 1: 2020-06-02 05:12:14.973 2020-06-02 06:24:51.367 1.210109 hours

rbindlist(difftime_diffunits)

# start_time                end_time      diff_time
# 1: 2020-06-02 05:12:14.652 2020-06-02 06:01:24.753 49.168350 mins
# 2: 2020-06-02 05:12:14.656 2020-06-02 06:02:25.045 50.173150 mins
# 3: 2020-06-02 05:12:14.973 2020-06-02 06:32:28.081  1.336974 mins
# 4: 2020-06-02 05:12:14.973 2020-06-02 06:24:51.367  1.210109 mins

Note that the units of diff_time of the 3rd and 4th rows should be "hours" instead of "mins".

# Output of sessionInfo()

R version 4.0.0 (2020-04-24)
Platform: x86_64-w64-mingw32/x64 (64-bit)
Running under: Windows Server x64 (build 17763)

Matrix products: default

locale:
[1] LC_COLLATE=English_United States.1252  LC_CTYPE=English_United States.1252    LC_MONETARY=English_United States.1252
[4] LC_NUMERIC=C                           LC_TIME=English_United States.1252    

attached base packages:
[1] stats     graphics  grDevices utils     datasets  methods   base     

other attached packages:
[1] data.table_1.12.9

loaded via a namespace (and not attached):
[1] compiler_4.0.0 tools_4.0.0    tinytex_0.23   xfun_0.14     

[jangorecki]

Thank you for reporting. The problem is quite tricky because attributes of diff time are defining actual value. We could make a special handling of such class, but it is not really addressing the problem because other classes that depends on attributes will still suffers from that. For example a decimal class where precision is stored in attribute.
I think it is better to just document that instead.

[CharnelMouse]

Simpler example using by:

library(data.table)

x <- data.table(a = c(1L, 1L, 2L, 2L),
                b = as.POSIXct(c("2020-07-03 12:00:00", "2020-07-03 12:00:01",
                                 "2020-07-03 12:00:00", "2020-08-05 12:00:00")))

x

#    a                   b
# 1: 1 2020-07-03 12:00:00
# 2: 1 2020-07-03 12:00:01
# 3: 2 2020-07-03 12:00:00
# 4: 2 2020-08-05 12:00:00

x[, .(b = difftime(b[2], b[1], units = "auto"), by = a]

#    a       b
# 1: 1  1 secs
# 2: 2 33 secs

x[, .(b = difftime(b[2], b[1], units = "secs")), by = a]

#    a            b
# 1: 1       1 secs
# 2: 2 2851200 secs