in dcast, with multiple fun.agg and a single value.var, don't put the var in the col names

Start by reproducing the dcast(melt(data, id = c("id", "trt")), variable ~ trt, fun = list(mean, sd)) example in the data.table package. Read the dcast documentation and its implementation to find how value.var and multiple aggregation functions determine output names. Done means the redundant value_ prefix is omitted while the requested mean and sd columns remain distinguishable.

This is a rehash of an earlier request, #1810

data <- data.frame(id = c(1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36,37,38,39,40,41,42,43,44,45,46,47,48),
                   x1 = c(0.2846,0.3741,0.4208,0.3756,0.3476,0.3664,0.2852,0.3537,0.3116,0.3124,0.364,0.3934,0.3456,0.3034,0.3139,0.2766,0.3034,0.3159,0.3648,0.4046,0.3961,0.3451,0.2059,0.3184,0.2481,0.3503,0.331,0.3166,0.3203,0.1868,0.245,0.1625,0.2227,0.196,0.1697,0.2064,0.1369,0.1938,0.1498,0.1315,0.1523,0.2151,0.168,0.1427,0.3083,0.301,0.2328,0.2747),
                   x2 = c(-0.4364,-0.5262,-0.5338,-0.5037,-0.4758,-0.5003,-0.4359,-0.5002,-0.4027,-0.424,-0.4811,-0.5492,-0.3846,-0.3899,-0.4473,-0.3688,-0.3946,-0.4112,-0.4833,-0.4909,-0.4865,-0.368,0.295,-0.3221,-0.2482,-0.5424,-0.5021,-0.4453,-0.3952,0.3915,0.4472,0.364,0.436,0.3877,0.4077,0.2737,0.3104,0.3514,0.3256,0.287,0.3126,0.3648,-0.2596,-0.1913,-0.3656,-0.4598,-0.3198,-0.3685),
                   x3 = c(0.6043,0.5141,0.4638,0.486,0.3691,0.4104,0.426,0.3846,0.3191,0.4347,0.5842,0.4638,0.4418,0.523,0.5009,0.4568,0.5105,0.5421,0.4857,0.4063,0.391,0.4114,0.5189,0.5248,0.4942,0.2855,0.6107,0.4712,0.2009,0.4632,0.4457,0.3914,0.4547,0.4801,0.4873,0.5501,0.4442,0.4458,0.4651,0.5748,0.5231,0.4869,0.1769,0.099,0.5013,0.4543,0.4601,0.4396),
                   x4 = c(0.4895,0.6991,0.6566,0.6106,0.6976,0.6883,0.6533,0.6951,0.6852,0.5062,0.5682,0.6172,0.5073,0.6514,0.577,0.5228,0.6571,0.6132,0.4893,0.7904,0.6519,0.6582,0.6919,0.6011,0.6145,0.5943,0.4608,0.5997,0.4431,0.4082,0.5641,0.4535,0.5448,0.4632,0.4237,0.6187,0.4115,0.4995,0.4504,0.4103,0.4511,0.527,0.3654,0.2537,0.6317,0.478,0.5915,0.5283),
                   trt = c("A","A","A","A","A","A","A","A","A","A","A","A","A","A","B","B","B","B","B","B","B","B","B","B","B","B","B","B","B","C","C","C","C","C","C","C","C","C","C","C","C","C","D","D","D","D","D","D")
                   )
library(data.table)
setDT(data)
dcast(melt(data, id = c("id", "trt")), variable ~ trt, fun = list(mean, sd))

   variable value_mean_A value_mean_B value_mean_C value_mean_D value_sd_A value_sd_B value_sd_C value_sd_D
1:       x1    0.3456000    0.3207333    0.1821923    0.2379167 0.04104517 0.05143110 0.03501070 0.06966645
2:       x2   -0.4674143   -0.3740600    0.3584308   -0.3274333 0.05489628 0.20033792 0.05489583 0.09365470
3:       x3    0.4589214    0.4540600    0.4778769    0.3552000 0.07952784 0.10363689 0.04866813 0.17130249
4:       x4    0.6232571    0.5976867    0.4789231    0.4747667 0.07624950 0.09146218 0.06686731 0.14280235

Since value_* is a prefix on all columns, it is redundant and I'd prefer to have it omitted.

Example from SO: https://stackoverflow.com/q/46246048/