Avoid paste0 overhead in is_lint_level

[MichaelChirico]

Closes #2358

[codecov-commenter]

Codecov Report

All modified and coverable lines are covered by tests ✅

Comparison is base (f6453b6) 99.13% compared to head (649b9f9) 99.13%.

❗ Current head 649b9f9 differs from pull request most recent head e61c5b0. Consider uploading reports for the commit e61c5b0 to get more accurate results

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #2360      +/-   ##
==========================================
- Coverage   99.13%   99.13%   -0.01%     
==========================================
  Files         124      124              
  Lines        5574     5573       -1     
==========================================
- Hits         5526     5525       -1     
  Misses         48       48              

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[AshesITR]

Can you benchmark switch() vs. if() first?

[IndrajeetPatil]

I had done a quick benchmarking locally that I am producing here, but not sure if you are actually looking for benchmark for lint_package() (e.g.) after this change. I didn't do the latter before approving.

level <- "expression"

bench::mark(
  "if" = paste0(if (level == "file") "full_", "parsed_content"),
  "switch" = switch(level,
    file = "full_parsed_content",
    expression = "parsed_content"
  ),
  iterations = 1000,
  time_unit = "us"
)
#> # A tibble: 2 × 6
#>   expression   min median `itr/sec` mem_alloc `gc/sec`
#>   <bch:expr> <dbl>  <dbl>     <dbl> <bch:byt>    <dbl>
#> 1 if         0.902 0.984    881984.        0B        0
#> 2 switch     0     0.0410 21351784.        0B        0

^{Created on 2023-11-28 with reprex v2.0.2}

[AshesITR]

I was thinking if (level == "expression") "parsed_content" else "full_parsed_content", still without paste0().

[IndrajeetPatil]

Ah, I see. Here it is:

level <- "expression"

bench::mark(
  "if" = if (level == "file")  "full_parsed_content" else "parsed_content",
  "switch" = switch(level,
                    file = "full_parsed_content",
                    expression = "parsed_content"
  ),
  iterations = 1000,
  time_unit = "us"
)
#> # A tibble: 2 × 6
#>   expression    min median `itr/sec` mem_alloc `gc/sec`
#>   <bch:expr>  <dbl>  <dbl>     <dbl> <bch:byt>    <dbl>
#> 1 if         0.0410 0.0820  6850171.        0B        0
#> 2 switch     0      0.0410 19969240.        0B        0

^{Created on 2023-11-28 with reprex v2.0.2}

[AshesITR]

On a computer now, could check,'

level <- "expression"

bench::mark(
  "if" = if (level == "expression") "parsed_content" else "full_parsed_content",
  "switch" = switch(level,
    file = "full_parsed_content",
    expression = "parsed_content"
  )
)
#> # A tibble: 2 × 6
#>   expression      min   median `itr/sec` mem_alloc `gc/sec`
#>   <bch:expr> <bch:tm> <bch:tm>     <dbl> <bch:byt>    <dbl>
#> 1 if            120ns    128ns  6010347.        0B        0
#> 2 switch         76ns     83ns  9321876.        0B        0

level <- "file"

bench::mark(
  "if" = if (level == "expression") "parsed_content" else "full_parsed_content",
  "switch" = switch(level,
    file = "full_parsed_content",
    expression = "parsed_content"
  )
)
#> # A tibble: 2 × 6
#>   expression      min   median `itr/sec` mem_alloc `gc/sec`
#>   <bch:expr> <bch:tm> <bch:tm>     <dbl> <bch:byt>    <dbl>
#> 1 if          123.1ns  132.1ns  6368058.        0B        0
#> 2 switch       68.9ns   75.1ns 10365642.        0B        0

^{Created on 2023-11-28 with reprex v2.0.2}

[AshesITR]

Here are some larger benchmarks:

expression

# A tibble: 2 × 13
  expression      min   median `itr/sec` mem_alloc `gc/sec`   n_itr  n_gc total_time result    memory             time                    gc                       
  <bch:expr> <bch:tm> <bch:tm>     <dbl> <bch:byt>    <dbl>   <int> <dbl>   <bch:tm> <list>    <list>             <list>                  <list>                   
1 if            115ns    129ns  6658306.        0B     17.3 9999974    26       1.5s <chr [1]> <Rprofmem [0 × 3]> <bench_tm [10,000,000]> <tibble [10,000,000 × 3]>
2 switch       70.9ns     80ns 10382848.        0B     26.0 9999975    25    963.1ms <chr [1]> <Rprofmem [0 × 3]> <bench_tm [10,000,000]> <tibble [10,000,000 × 3]>

file

# A tibble: 2 × 13
  expression      min   median `itr/sec` mem_alloc `gc/sec`   n_itr  n_gc total_time result    memory             time                    gc                       
  <bch:expr> <bch:tm> <bch:tm>     <dbl> <bch:byt>    <dbl>   <int> <dbl>   <bch:tm> <list>    <list>             <list>                  <list>                   
1 if            118ns  134.1ns  6591956.        0B     16.5 9999975    25      1.52s <chr [1]> <Rprofmem [0 × 3]> <bench_tm [10,000,000]> <tibble [10,000,000 × 3]>
2 switch         63ns   70.1ns 11524237.        0B     28.8 9999975    25   867.73ms <chr [1]> <Rprofmem [0 × 3]> <bench_tm [10,000,000]> <tibble [10,000,000 × 3]>