merge diagnostics argument/option

[grantmcdermott]

Prompted by this tweet by @johnjosephhorton (whom, I hope, doesn't mind being tagged).

An R function I'd find useful: diagnostics of tables joins---column overlap to start, multiple matches, # of non-matches, missing keys & how they were handled, maybe a venn diagram of column names, etc. Something like this already exist? Joining is a high risk data cleaning task

I'd advocate for a new diagnostics logical argument (whose default behaviour could fall back to an option), since verbose already returns quite a lot of unrelated information.

Worth noting that the tidylog package largely provides this behaviour for dplyr joins.

library(dplyr, warn.conflicts = FALSE)
library(tidylog, warn.conflicts = FALSE)

x = data.frame(A = 1:5, B = 6:10)
y = data.frame(A = c(1, 4), C = LETTERS[c(1, 4)])

left_join(x, y)
#> Joining, by = "A"
#> left_join: added one column (C)
#>            > rows only in x   3
#>            > rows only in y  (0)
#>            > matched rows     2
#>            >                 ===
#>            > rows total       5
#>   A  B    C
#> 1 1  6    A
#> 2 2  7 <NA>
#> 3 3  8 <NA>
#> 4 4  9    D
#> 5 5 10 <NA>

^{Created on 2021-12-27 by the reprex package (v2.0.1)}

As per John's tweet, additional data.table-specific information like keys would be valuable here too. Doesn't resolve, but would go quite far towards heading off issues like #4888 and #4891.

Session Info

sessionInfo()
#> R version 4.1.2 (2021-11-01)
#> Platform: x86_64-pc-linux-gnu (64-bit)
#> Running under: Arch Linux
#> 
#> Matrix products: default
#> BLAS/LAPACK: /usr/lib/libopenblas_haswellp-r0.3.18.so
#> 
#> locale:
#>  [1] LC_CTYPE=en_US.UTF-8       LC_NUMERIC=C              
#>  [3] LC_TIME=en_US.UTF-8        LC_COLLATE=en_US.UTF-8    
#>  [5] LC_MONETARY=en_US.UTF-8    LC_MESSAGES=en_US.UTF-8   
#>  [7] LC_PAPER=en_US.UTF-8       LC_NAME=C                 
#>  [9] LC_ADDRESS=C               LC_TELEPHONE=C            
#> [11] LC_MEASUREMENT=en_US.UTF-8 LC_IDENTIFICATION=C       
#> 
#> attached base packages:
#> [1] stats     graphics  grDevices utils     datasets  methods   base     
#> 
#> other attached packages:
#> [1] tidylog_1.0.2 dplyr_1.0.7  
#> 
#> loaded via a namespace (and not attached):
#>  [1] pillar_1.6.4      compiler_4.1.2    highr_0.9         R.methodsS3_1.8.1
#>  [5] R.utils_2.11.0    tools_4.1.2       digest_0.6.29     evaluate_0.14    
#>  [9] lifecycle_1.0.1   tibble_3.1.6      R.cache_0.15.0    pkgconfig_2.0.3  
#> [13] rlang_0.4.12      reprex_2.0.1      DBI_1.1.1         yaml_2.2.1       
#> [17] xfun_0.28         fastmap_1.1.0     withr_2.4.3       styler_1.6.2     
#> [21] stringr_1.4.0     knitr_1.36        generics_0.1.1    fs_1.5.2         
#> [25] vctrs_0.3.8       tidyselect_1.1.1  glue_1.5.1        R6_2.5.1         
#> [29] fansi_0.5.0       rmarkdown_2.11    tidyr_1.1.4       purrr_0.3.4      
#> [33] magrittr_2.0.1    clisymbols_1.2.0  backports_1.4.1   ellipsis_0.3.2   
#> [37] htmltools_0.5.2   assertthat_0.2.1  utf8_1.2.2        stringi_1.7.6    
#> [41] crayon_1.4.2      R.oo_1.24.0

^{Created on 2021-12-27 by the reprex package (v2.0.1)}

[ben-schwen]

Related/potential duplicate to #4677

[grantmcdermott]

Ah, don't know how I missed that. Agree it's a dup, so will close and add a comment there.