GroupBy with an array in dataviewjs

[awlobo]

Hi!
I want to group by a yaml property which is an array with dataviewjs.
This is an example of the header of my files:

---
name: string
url: string
location: string
type: [sushi, japanese, vegetarian]
---

Doing let groups = dv.pages("#restaurant").groupBy(p => p.type) returns Evaluation Error: RangeError: Maximum call stack size exceeded and I also tried with let groups = dv.pages("#restaurant").groupBy(p => p.type.values) but the grouping is not correct as it adds x restaurant to a group with a typenot included in that restaurant.

In a nutshell, I want something similar to:

TABLE WITHOUT ID type, rows.file.link
FROM #restaurant 
FLATTEN type
GROUP BY type

but with this structure:

## sushi

Table with restaurants with sushi type

## japanese

Table with restaurants with japanese type

## vegetarian

Table with restaurants with vegetarian type

etc.

[blacksmithgu]

Hmm, I need a flatten operator for data arrays it seems. You can hack it for now like this:

let groups = dv.pages("#restaurant").flatMap(page => page.type.map(t => Object.assign({}, page, { type: t })))
    .groupBy(page => page.type);

[HyperEpsilon]

I believe it has something to do with how DataView stores multiple key value pairs. Or more specifically, how they store a single value as a value and multiple values as a list.

To make it safe, I would need to check if type is a list, and if not then make it a list (possibly empty) and go from there. And also check to ensure that type is even a member of the object.

[AB1908]

Strange. Multiple defs of the same key in a note is indexed as an array.

[HyperEpsilon]

I should have specified. If a note 1 has type:: A and type::B in the text, then type is considered an array (which is what I want). However, if note 2 only has type:: b in the text, then that is not considered an array and flatMap does not work. The same is true if in the frontmatter the key is defined as type: a rather than type: [a].

[AB1908]

Oh fair fair. Use dv.array to wrap the field in an array.

[HyperEpsilon]

dv.array works perfectly! Here is my completed version:

let groups = dv.pages("#restaurant").flatMap(page => dv.array(page.type).map(t => Object.assign({}, page, { type: t })))
    .groupBy(page => page.type);

for (let group of groups) {
	dv.header(3, group.key);
	dv.table(["Name", "Type"],
		group.rows
			.map(k => [k.file.link, k.type])
		)
}

Tested using all different types of data annotation: frontmatter single values, frontmatter arrays, frontmatter lists, inline single values, inline multiple values, and non-existent.

[HyperEpsilon]

It was buried in the replies of a comment, but here is a version that I have that works on all forms of data annotation.

let groups = dv.pages("#restaurant").flatMap(page => dv.array(page.type).map(t => Object.assign({}, page, { type: t })))
    .groupBy(page => page.type);

for (let group of groups) {
	dv.header(3, group.key);
	dv.table(["Name", "Type"],
		group.rows
			.map(k => [k.file.link, k.type])
		)
}