sep argument ignored in dcast

[RoyalTS]

dcast seems to be ignoring the sep argument:

dt <- data.table(user=rep(1:10, each=2),
                 measurement=rep(c('a', 'b'), times=10),
                 value=rnorm(20))

dcast(dt, user~measurement, sep='_')

This yields

    user          a           b
 1:    1 -1.1240899 -1.00620808
 2:    2  0.3236956 -1.86065046
 3:    3 -0.9664502  0.21878106
 4:    4  0.9231289 -0.86210577
 5:    5  1.8197930 -0.51468111
 6:    6  0.5600921 -0.64247822
 7:    7  1.0440452 -0.03289074
 8:    8 -0.1192868 -0.79392246
 9:    9  1.0655679 -1.83930778
10:   10  0.2463848  0.43551250

Given _ as the sep – or indeed if I had left the sep argument blank – the column names should be measurement_a and measurement_b, no?

> sessionInfo()
R version 3.3.2 (2016-10-31)
Platform: x86_64-apple-darwin13.4.0 (64-bit)
Running under: macOS Sierra 10.12.4

locale:
[1] en_US.UTF-8/en_US.UTF-8/en_US.UTF-8/C/en_US.UTF-8/en_US.UTF-8

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

other attached packages:
[1] data.table_1.10.4

loaded via a namespace (and not attached):
[1] tools_3.3.2

[MichaelChirico]

This is expected behavior, but perhaps ?dcast could be more of a help here. sep is used when you have multiple value.var:

dt <- data.table(user=rep(1:10, each=2),
                 measurement=rep(c('a', 'b'), times=10),
                 value=rnorm(20),
                 value2 = rnorm(20))
dcast(dt, user~measurement, sep='_', value.var = c('value', 'value2'))
#     user     value_a    value_b    value2_a     value2_b
#  1:    1  2.47824187 -0.4655328 -1.81186160 -0.295153290
#  2:    2  0.27597669  0.0499813  1.09270848  1.465269972
#  3:    3 -0.51353846 -1.4100579  1.65814414  1.841348213
#  4:    4 -1.14161843  0.3337608 -0.05070586  0.287335488
#  5:    5 -1.11248648 -0.2373563  0.72059394 -1.176779744
#  6:    6 -0.74958142 -0.4015603  2.47068888 -0.759265311
#  7:    7  0.22767623 -1.6260782  1.36925862  0.476299662
#  8:    8  0.61496455  1.6415110 -0.72233829 -0.003333061
#  9:    9  0.51846450  1.1448276  0.16129185  0.444361567
# 10:   10  0.05730503  1.1189461  1.34467875 -0.477063731

[RoyalTS]

The case in which variables are named value_a and value_b while in wide form with the separator being _ or any other separator seems to me to be super common. That there's no easy way to produce that wide format from a long-format table when dcast has a sep argument is super confusing.

So yeah, some clarification in the docs at the very least would be great.

[franknarf1]

sep is used when you have multiple value.var

@MichaelChirico It is also used to separate multiple functions. Probably related:

• FR: in dcast, with multiple value.var and a single fun.agg, don't put the fun in the col names #1810
• FR: in dcast, with g ~ ., don't put . in the col names #1821
• [question]dcast with multiple value.var : order between measure.vars and value.var #1951 (comment)

[arunsrinivasan]

The behaviour here is consistent (as explained above). If you'd like to issue a PR reflecting the necessary changes to help file, it'd be great!