[ARROW-16039] [JS] Documentation is quite obscure and not useful

• Type: Improvement
• Status: Open
• Priority: Major
• Resolution: Unresolved
• Affects Version/s: 7.0.0
• Component/s: JavaScript
• Environment: Linux, Deno, DOM
• Flags: Important
• External issue URL: https://github.com/apache/arrow/issues/31458

Language

javascript

• Assignee: Unassigned
• Reporter: Teodor Kostov
• Created: 27/Mar/22 18:31
• Updated: 11/Jan/23 11:41

Description

I've been looking forward to using Apache Arrow as a data storage component in a frontend application that heavily relies on time series data. However, the syntax seems to have changed quite a lot with version 7.0.0. Most of the examples on https://observablehq.com seem to be outdated because of that.

The two main resources https://arrow.apache.org/docs/js/index.html and https://arrow.apache.org/docs/js/modules/Arrow_dom.html are quite insufficient to understand how to use the project. There are a bunch of examples on how to create a table and a vector. However, it seems that the most important use case in JS is not captured - how to create a table from an array of records

[{a:1, b:2}, {a:3, b:4}]

or how to create a table from an observable that provides one record at a time.

callback(record => ???)

No information on how to append data to a table (except the concat() method).

No information on how to manipulate the data in a table or an example on how to consume it beyond the fact that a table is an iterable and has a get() method.

Please, it will be also quite helpful to add some examples on how to work with time series data.

Current state of the documentation does not provide the ability for the project to be adopted by anyone else except the core developers.

Issue Links

• is a parent of: ARROW-16153 [JS] Consider implementing a tableFromArray
• is duplicated by: ARROW-12531 [JS] Make the docs more user friendly