How to group tasks by File and then by Section?

[claremacrae]

I've been using this a lot, to ask dataview to get all the tasks from a set of files:

TASK
FROM #current
WHERE !fullyCompleted
SORT file.name

But I would like to extend this to group the tasks by section heading...

So something like this:

## File 1

### Section 1

- [ ] Task 1

### Section 2

- [ ] Task 2

## File 2

### Section 1

- [ ] Task 1

### Section 2

- [ ] Task 2

I know that tasks now contain sections, but I cannot for the life of me work out how to use two levels of groupings.

Any suggestions please?

---

Edit - 2023-08-13

The two answers that produce two levels of groups (filename, then heading), using purely dataview/dataviewjs are:

• this comment by @claremacrae
• this comment by @ces3001

[MugenLee]

I want to get it to
In the Official document

it seen to can use "meta(link).subpath" to accomplish that display tasks under specific headings

But I have not been able to achieve: display the task under the specified title
Any suggestions please?

[jesvtb]

it seems that the the subpath is case sensitive. I tried with mine and the code block example from official doc works fine.

[claremacrae]

Thanks.

This is the best I was able to manage - using a dataviewjs block:

function showNotFullyCompleteTasksGroupedBySection(page) {
    let all_tasks = page.file.tasks.where(t => !t.fullyCompleted)
    if (all_tasks.length == 0)
    {
        return
    }

    dv.header(2, page.file.name)
    
    // TODO this lists the sections in alphabetical order,
    //      but I wish to see them in the order they are
    //      in the file
    for (let group of all_tasks.groupBy(t => t.section)) {
        // This header shows a link labelled "File name > Section Name"
        // TODO I would like it to show just "Section Name" in the hyperlink
        dv.header(3, group.key)
        dv.taskList(group.rows, false)
    }
}

for (let page of dv.pages().sort(p => p.file.name, 'asc')) {
    showNotFullyCompleteTasksGroupedBySection(page)
}

Some example output:

[MugenLee]

Thanks.

This is the best I was able to manage - using a dataviewjs block:

function showNotFullyCompleteTasksGroupedBySection(page) {
    let all_tasks = page.file.tasks.where(t => !t.fullyCompleted)
    if (all_tasks.length == 0)
    {
        return
    }

    dv.header(2, page.file.name)
    
    // TODO this lists the sections in alphabetical order,
    //      but I wish to see them in the order they are
    //      in the file
    for (let group of all_tasks.groupBy(t => t.section)) {
        // This header shows a link labelled "File name > Section Name"
        // TODO I would like it to show just "Section Name" in the hyperlink
        dv.header(3, group.key)
        dv.taskList(group.rows, false)
    }
}

for (let page of dv.pages().sort(p => p.file.name, 'asc')) {
    showNotFullyCompleteTasksGroupedBySection(page)
}

Thank you for the method
I achieved it in a more familiar way, for reference~

```dataview
task 

where 
header = [[file.name#header]]

```

[senged]

Thanks for the js example. I tried it but it gives all the content around the task up to the header it seems. Have you made any updates to this since your last post? (I'm very green to the dataviewjs, not yet saavy to try to edit it)

[AB1908]

@senged you may want to hold off on the JS part for a while. Sections will be treated as first class citizens in an upcoming release.

[claremacrae]

Thank you for the method I achieved it in a more familiar way, for reference~

```dataview
task 

where 
header = [[file.name#header]]

Thanks. But this does not do either the search or the 2-level grouping that I was asking about in the original post.

[angelv56]

Thanks for your anwers guys, I just recent see this post to do the same that you want but also did not find a solution that would allow me to solve the TODOs that you comment, after a while I found the next solution, first you must have the obsidian-tasks plugin installed and then with the following code you can use a combination of dataviewjs with the tasks plugin syntax to perform the grouped as you require, without showing the name of the file and the section thogether, also you have the power that the same plugin offers to be able to edit the tasks and other benefits, in this example i am showing the pending tasks, filtering and sorting by the "type" and "priority" metadata but you can replace by a tag or any other filter, for more references check this link

```dataviewjs
function listTasks(page){
	const query = `
	not done
	path includes ${page.file.path}
	# you can add any number of extra Tasks instructions, for example:
	group by heading
	`;
		
	dv.el("h3", page.file.name);
	dv.paragraph('```tasks\n' + query + '\n```');
}

for (let page of dv.pages().where(t => t.type == "pendings").sort(p => p.priority, 'asc')) {
    listTasks(page)
}
```

[AB1908]

Very interesting interop with the tasks plugin. I'm amazed.

[claremacrae]

Thanks. As the maintainer and current author of the Tasks plugin, I'm pleased to see it being used.

One thing about this solution is that it if any files match the dv.pages() filter but don't have any tasks in them, what is written out is the file name and an empty tasks block.

I really was looking for a pure dataview solution, so I will keep looking through the answers.

[randompherret]

For whoever gets here from google. You can also just do something like

TASK
FROM #current
WHERE !fullyCompleted
GROUP BY header

The header starts with the filename, so it will naturally sort by file and then section.

[claremacrae]

Thanks, but if you look at the requested output in the first post, it shows I was requesting two levels of grouping not one.

[acdesigns]

So there's no way using regular dataview to do this? I'm not comfortable using js and would rather keep is really simple. I'm trying to do something similar.

I have a task list in the top of the note containing the tasks and just want them separated by header ONLY, without the name of the file.

Is there nothing I can do without js? Thx

[acdesigns]

I found a way!
Used this to help me figure it out:
Obsidian Forum: task-list-of-a-specific-section-of-all-daily-notes

Here's the final code:
`

TASK
WHERE file.name = this.file.name
WHERE status = "?"
GROUP BY meta(section).subpath

`

Hope it helps someone else, It MADE MY DAY :)

[claremacrae]

Thanks, but if you look at the requested output in the first post, it shows I was requesting two levels of grouping not one.

This groups by the heading level, but it does not first group by the filename.

[ces3001]

This should work to group and sort the tasks by order of appearance of the sections and not alphabetically.
taskList can contain tasks from multiple files as well, just change the first taskList = line to get the tasks you want.

	```dataviewjs
	const summary = true
	const includeSection = true
	let taskList = dv.current().file.tasks
				 .where(t => !t.completed && dv.date(t.start) <= dv.date('today'))
	taskList = taskList
		.sort(t => t.line)
		.sort(t => t.path)
	if (summary) { dv.span("*" + taskList.length + " tasks*\n") }
	if (includeSection) {
		const regex = /> ([^\]]+)/;
		let section = null
		let page = null
		let sectionTaskList = []
		for (let t of taskList) {
			if ((sectionTaskList.length > 0) && 
				((page != t.path) || (section != t.section.toString()))) {
				dv.taskList(sectionTaskList,false)
			}
			if (page != t.path) {
				dv.header(2,t.path.replace(/\.md$/, ''))
				page = t.path
			}
			if (section != t.section.toString()) {
				section = t.section.toString()
				dv.header(3,  section.match(regex)[1])
				sectionTaskList = []
			}
			sectionTaskList = sectionTaskList.concat(t)
		}
		dv.taskList(sectionTaskList,false)
	} else {
		dv.taskList(taskList,true)  // true = divide by file
	}
	``` 

[claremacrae]

Thanks for the reply.
And wow - this works amazingly well! 🎉

Here is what it looks like, as written, but with dv.pages("#adobe/lightroom"):

In the where clause, I always prefer to use t.fullyCompleted rather than t.completed, to see the full picture. That gives this:

And I really like the fact that I can reduce the volume of output by turning the true to false on the first two lines, to get this:

[timespaced]

This is great, and exactly what I was looking for as well!

I'll note, though, if you have a task at the top of a file with no section/header, the script errors out. I don't know crap about coding - any chance you might be able to put in a failsafe for this?

[ces3001]

This is great, and exactly what I was looking for as well!

I'll note, though, if you have a task at the top of a file with no section/header, the script errors out. I don't know crap about coding - any chance you might be able to put in a failsafe for this?

Try this:
You can replace the "No section” for whatever you want, or even remove the whole line if you prefer no separation.

	const summary = true
	const includeSection = true
	let taskList = dv.current().file.tasks
				 .where(t => !t.completed && dv.date(t.start) <= dv.date('today'))
	taskList = taskList
		.sort(t => t.line)
		.sort(t => t.path)
	if (summary) { dv.span("*" + taskList.length + " tasks*\n") }
	if (includeSection) {
		const regex = /> ([^\]]+)/;
		let section = null
		let page = null
		let sectionTaskList = []
		for (let t of taskList) {
			if ((sectionTaskList.length > 0) && 
				((page != t.path) || (section != t.section.toString()))) {
				dv.taskList(sectionTaskList,false)
			}
			if (page != t.path) {
				dv.header(2,t.path.replace(/\.md$/, ''))
				page = t.path
			}
			if (section != t.section.toString()) {
				section = t.section.toString()
				if (section.match(regex) != null) {
					dv.header(3,  section.match(regex)[1]) 
				} else {
					dv.header(3, "No section")
				}
				sectionTaskList = []
			}
			sectionTaskList = sectionTaskList.concat(t)
		}
		dv.taskList(sectionTaskList,false)
	} else {
		dv.taskList(taskList,true)  // true = divide by file
	}