[stdlib-candidate] add aggregate

[Bahex]

Description

Add aggregate, a command that operates on the output of group-by --to-table to help aggregate to do quick inspections.

Related

• add agg-by as a command that helps aggregate data with sum, avg, and count nushell#14316 (comment)
• A more sql-like group-by nushell#2607
• add multiple grouper support to group-by nushell#14337

Examples

open ~/Downloads/movies.csv
  | group-by Lead_Studio Genre --to-table
  | aggregate Worldwide_Gross
  # | first 4
  # | to md

Lead_Studio	Genre	count	Worldwide_Gross_min	Worldwide_Gross_avg	Worldwide_Gross_max	Worldwide_Gross_sum
The Weinstein Company	Comedy	1	19.62	19.62	19.62	19.62
The Weinstein Company	Drama	1	8.26	8.26	8.26	8.26
Independent	Comedy	7	14.31	57.01	205.3	399.07
Independent	Romance	7	0.03	149.82142857142858	702.17	1048.75

---

open ~/Downloads/movies.csv
  | group-by Lead_Studio Genre --to-table
  | aggregate Worldwide_Gross --ops {avg: {math avg}, std: {math stddev}}
  # | first 4
  # | to md

Lead_Studio	Genre	count	Worldwide_Gross_avg	Worldwide_Gross_std
The Weinstein Company	Comedy	1	19.62	0
The Weinstein Company	Drama	1	8.26	0
Independent	Comedy	7	57.01	66.1709932134704
Independent	Romance	7	149.82142857142858	229.79475832816996

---

open ~/Downloads/movies.csv
  | group-by Lead_Studio Genre --to-table
  | aggregate Worldwide_Gross Audience_score_% --ops {avg: {math avg}}
  # | first 4
  # | to md

Lead_Studio	Genre	count	Worldwide_Gross_avg	Audience_score_%_avg
The Weinstein Company	Comedy	1	19.62	52
The Weinstein Company	Drama	1	8.26	84
Independent	Comedy	7	57.01	60.142857142857146
Independent	Romance	7	149.82142857142858	59.857142857142854

[fdncred]

My initial thought is to just land this PR because I want to use it. But we need to follow the readme so we need tests. The documentation is probably ok but I'd like to see more examples with different agg ops.

I love this script, and I already have a version of it from discord in my own custom commands that I source at every startup. Good work and thanks for adding this here.

[fdncred]

Thanks! Let's go!

[NotTheDr01ds]

@Bahex I'd like to propose that we promote this into std, but two things are holding me back:

1. There's now (on Nushell main) a failing test which I believe is due to the changes in Add run-time type checking for command pipeline input nushell#14741 - I'm not sure how you want to handle that. Previously, you special-cased a helpful error message suggesting group-by --to-table, but I don't think that's possible after the Add run-time type checking for command pipeline input nushell#14741 changes.
2. Perhaps less important, but at the moment this should probably be in std/tables/aggregate.nu. This will match up with the other PR I still need to move over from the "old" std-rfc. However, the current std implementation doesn't support more than one "nesting". I'd really like to fix that before we promote this, but it may take me a bit.

[Bahex]

@NotTheDr01ds

1. While not ideal, I can add a second input output type of record -> error to keep this helpful suggestion? I checked out the current error message and depending on the input record the error can contain an extremely input type.

Details