merge.data.table ignores incomparables without warnings.

[GBsuperman]

example of using 'incomparables' in DF.

x <- data.frame(k1 = c(NA,NA,3,4,5), k2 = c(1,NA,NA,4,5), data = 1:5)
y <- data.frame(k1 = c(NA,2,NA,4,5), k2 = c(NA,NA,3,4,5), data = 1:5)
merge(x, y, by = "k2", incomparables = NA) # 2 rows
#   k2 k1.x data.x k1.y data.y
# 1  4    4      4    4      4
# 2  5    5      5    5      5

example of using 'incomparables' in DT.

x <- data.table(k1 = c(NA,NA,3,4,5), k2 = c(1,NA,NA,4,5), data = 1:5)
y <- data.table(k1 = c(NA,2,NA,4,5), k2 = c(NA,NA,3,4,5), data = 1:5)
merge(x, y, by = "k2", incomparables = NA) # 6 rows
#    k2 k1.x data.x k1.y data.y
# 1: NA   NA      2   NA      1
# 2: NA   NA      2    2      2
# 3: NA    3      3   NA      1
# 4: NA    3      3    2      2
# 5:  4    4      4    4      4
# 6:  5    5      5    5      5

Output of sessionInfo()

R version 3.3.3 (2017-03-06)
Platform: x86_64-apple-darwin13.4.0 (64-bit)
Running under: macOS  10.13.1

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] parallel  stats4    stats     graphics  grDevices utils     datasets  methods   base     

other attached packages:
 [1] bindrcpp_0.2          tidyr_0.7.2           dplyr_0.7.4           ggplot2_2.2.1        
 [5] gtools_3.5.0          knitcitations_1.0.8   knitr_1.18            GBsuperman_0.0.0.9000
 [9] data.table_1.10.4-3   roxygen2_6.0.1        devtools_1.13.4       S4Vectors_0.12.2     
[13] BiocGenerics_0.20.0  

loaded via a namespace (and not attached):
 [1] nlme_3.1-131              bitops_1.0-6              lubridate_1.7.1           dimRed_0.1.0             
 [5] RColorBrewer_1.1-2        httr_1.3.1                rprojroot_1.3-2           grpreg_3.1-2             
 [9] ggbiplot_0.55             tools_3.3.3               backports_1.1.2           utf8_1.1.3               
[13] R6_2.2.2                  rpart_4.1-11              lazyeval_0.2.1            colorspace_1.3-2         
[17] nnet_7.3-12               withr_2.1.1               tidyselect_0.2.3          gridExtra_2.3            
[21] mnormt_1.5-5              curl_3.1                  git2r_0.21.0              cli_1.0.0                
[25] rvest_0.3.2               glmnet_2.0-13             xml2_1.1.1                desc_1.1.1               
[29] scales_0.5.0              sfsmisc_1.1-1             DEoptimR_1.0-8            psych_1.7.8              
[33] robustbase_0.92-8         commonmark_1.4            stringr_1.2.0             digest_0.6.13            
[37] foreign_0.8-69            pkgconfig_2.0.1           bibtex_0.4.2              highr_0.6                
[41] rlang_0.1.6               ddalpha_1.3.1             rstudioapi_0.7            bindr_0.1                
[45] jsonlite_1.5              ModelMetrics_1.1.0        RCurl_1.95-4.10           magrittr_1.5             
[49] leaps_3.0                 Matrix_1.2-12             Rcpp_0.12.14              munsell_0.4.3            
[53] RefManageR_0.14.20        stringi_1.1.6             pROC_1.10.0               ggpmisc_0.2.16           
[57] MASS_7.3-48               plyr_1.8.4                recipes_0.1.1             grid_3.3.3               
[61] crayon_1.3.4              lattice_0.20-35           splines_3.3.3             HiClimR_1.2.3            
[65] MetaboAnalystR_0.0.0.9000 pillar_1.0.1              webchem_0.3.0             reshape2_1.4.3           
[69] codetools_0.2-15          CVST_0.2-1                glue_1.2.0                evaluate_0.10.1          
[73] foreach_1.4.4             gtable_0.2.0              purrr_0.2.4               kernlab_0.9-25           
[77] assertthat_0.2.0          DRR_0.0.3                 gower_0.1.2               prodlim_1.6.1            
[81] broom_0.4.3               class_7.3-14              survival_2.41-3           timeDate_3042.101        
[85] RcppRoll_0.2.2            tibble_1.4.1              iterators_1.0.9           memoise_1.1.0            
[89] lava_1.5.1                bestglm_0.36              caret_6.0-78              ipred_0.9-6  

[MarkusBonsch]

good point. 2 Options. Implement "incomparables" or warn. I will see if implementing seems feasible on short notice. Otherwise, I will include the warning. Definitely, there will be a warning, if any argument is given that ends up in the ignored ....

[MarkusBonsch]

I have looked into the issue. It can be solved with reasonable effort, but while I was writing the corresponding unit tests, I stumbled upon a much bigger problem: #2592. The latter needs to be solved first, definitely. I have created a master issue to collect all open inconsistencies between merge.data.frame and merge.data.table: #2593.

[jaapwalhout]

StackOverflow answer to be updated when this is implemented.