add math min and math max to bench command

[fdncred]

Description

This PR adds min and max to the bench command.

❯ use std bench
❯ bench { dply -c 'parquet("./data.parquet") | group_by(year) | summarize(count = n(), sum = sum(geo_count)) | show()' | complete | null } --rounds 100 --verbose
100 / 100
╭───────┬───────────────────╮
│ mean  │ 71ms 358µs 850ns  │
│ min   │ 66ms 457µs 583ns  │
│ max   │ 120ms 338µs 167ns │
│ std   │ 6ms 553µs 949ns   │
│ times │ [list 100 items]  │
╰───────┴───────────────────╯

User-Facing Changes

Tests + Formatting

After Submitting

[maxim-uvarov]

I'm not sure why into int | into float is there. Maybe it is historical artifact that isn't necessary animore

timeit { do $code } | into int | into float

Without it we can get rid of from ns

mean: ($times | math avg | from ns)

[amtoine]

@maxim-uvarov
into int is to convert the duration given by timeit to the corresponding number of nanoseconds and into float is here to help the math stddev not overflow 😮

for instance

seq 1 100 | each { random int } | math stddev

Error: nu::shell::operator_overflow

  × Operator overflow.
   ╭─[entry #2:1:13]
 1 │ seq 1 100 | each { random int } | math stddev
   ·             ──┬─
   ·               ╰── multiply operation overflowed
   ╰────
  help: Consider using floating point values for increased range by promoting operand with 'into float'. Note: float has reduced precision!

vs

seq 1 100 | each { random int } | into float | math stddev

2579795210688796000

[maxim-uvarov]

@amtoine got it, thanks

[fdncred]

@amtoine are you good with these additions? I think min and max are pretty helpful here.

[amtoine]

@amtoine are you good with these additions? I think min and max are pretty helpful here.

yup, it's fine because mean and std are already there 👍

to have my full opinion, i think i would have only selected times if i were to write this command again nowadays, letting the user compute statistics at will 🤣
because, this begs the question: why not add the math median, math variance or even compute the MAD indicator? 😉

[fdncred]

true. let's move ahead with this. it's just software, we can change it if we want later.