t.subtasks.values.filter not working any more

[wenlzhang]

t.subtasks.values.filter does not work any more with Obsidian v0.15.1 and Dataview 0.5.34. Is there any change in the code base regarding this function?

Specifically, it gives the following errors:

Evaluation Error: TypeError: t.subtasks.values.filter is not a function
    at hideCompletedSubtasks (eval at <anonymous> (plugin:dataview), <anonymous>:12:30)
    at Array.map (<anonymous>)
    at eval (eval at <anonymous> (plugin:dataview), <anonymous>:41:35)
    at DataviewInlineApi.eval (plugin:dataview:19514:16)
    at evalInContext (plugin:dataview:19515:7)
    at asyncEvalInContext (plugin:dataview:19525:32)
    at DataviewJSRenderer.render (plugin:dataview:19546:19)
    at DataviewJSRenderer.onload (plugin:dataview:19153:14)
    at DataviewJSRenderer.e.load (app://obsidian.md/app.js:1:730639)
    at t.e.addChild (app://obsidian.md/app.js:1:731063)

An example of the complete DataviewJS code can be found as follows:

#771 (comment)

[AB1908]

Don't think you need .values TBH.

[blacksmithgu]

The subtasks field on a task has been kept around for legacy purposes to allow people to easily still filter based on subtask state, but it is not respected at all when rendering the actual task view.

[wenlzhang]

Thanks for the replies!

I tried to simplify the code and hide all completed tasks and sub-tasks when rendering them. However, the first code block below gives the following errors. This is probably caused by const hideCompletedTasks part but I did not figure out how to resolve this issue.

Evaluation Error: TypeError: t.filter is not a function
    at hideCompletedTasks (eval at <anonymous> (plugin:dataview), <anonymous>:3:4)
    at Array.map (<anonymous>)
    at eval (eval at <anonymous> (plugin:dataview), <anonymous>:16:35)
    at DataviewInlineApi.eval (plugin:dataview:19514:16)
    at evalInContext (plugin:dataview:19515:7)
    at asyncEvalInContext (plugin:dataview:19525:32)
    at DataviewJSRenderer.render (plugin:dataview:19546:19)
    at DataviewRefreshableRenderer.maybeRefresh (plugin:dataview:19147:22)
    at t.e.tryTrigger (app://obsidian.md/app.js:1:963726)
    at t.e.trigger (app://obsidian.md/app.js:1:963659)

// Define function for hiding completed tasks
const hideCompletedTasks = (t) => {
	t.filter(!t.checked || t.text.includes("[[")).map(hideCompletedSubtasks)
}  // Show only uncompleted tasks and items that contain note links


// Search for tasks
const allTasks = dv.current()
	.file
	.tasks
	.where(t => !t.checked)
	.sort(t => [t.due, t.p, t.i])


// Hide completed tasks
allTasks.values = allTasks.values.map(hideCompletedTasks)


// Show tasks as a list
dv.header(5, "==Actions in Current Note==: " + allTasks.length + " tasks")

// dv.el("p", allTasks.length)  // Task count

for (let group of allTasks
	 .groupBy(t => t.p)
	 .sort(group => group.key, 'asc')) {
		dv.header(5, group.key)
		// dv.header(2, group.key.file.path)
		dv.taskList(group.rows
			.sort(t => [t.c, t.p, t.i])
			, true)
}

I also tried the following code but it shows all tasks, including those checked and transferred.

dv.taskList(dv.current().file.tasks.where(t => !t.checked))

[AB1908]

You're using values again, in allTasks.

[wenlzhang]

You're using values again, in allTasks.

I tried to replace allTasks.values = allTasks.values.map(hideCompletedSubtasks) with allTasks = allTasks.map(hideCompletedSubtasks) and it gives the following errors:

Evaluation Error: TypeError: Assignment to constant variable.
    at eval (eval at <anonymous> (plugin:dataview), <anonymous>:26:10)
    at DataviewInlineApi.eval (plugin:dataview:19514:16)
    at evalInContext (plugin:dataview:19515:7)
    at asyncEvalInContext (plugin:dataview:19525:32)
    at DataviewJSRenderer.render (plugin:dataview:19546:19)
    at DataviewJSRenderer.onload (plugin:dataview:19153:14)
    at DataviewJSRenderer.e.load (app://obsidian.md/app.js:1:730639)
    at t.e.addChild (app://obsidian.md/app.js:1:731063)
    at t.addChild (app://obsidian.md/app.js:1:1634575)
    at Object.addChild (app://obsidian.md/app.js:1:1633600)

[AnnaKornfeldSimpson]

When you first use allTasks you mark it const. Change const allTasks to let allTasks. That will avoid the error above.

[wenlzhang]

After replacing t.subtasks with t.children, I almost get what I wanted, just with some confusion now.

The original tasks and sub-tasks are as follows:

- [ ] Task 1 - Uncompleted
    - Item 1.1
        - Item 1.1.1
            - [ ] Task 1.1.1.1 - Uncompleted
            - [x] Task 1.1.1.2 - Completed
        - [x] Task 1.2 - Completed
            - [ ] Task 1.2.1 - Uncompleted
            - Item 1.2.2
    - [x] Task 1.2 - Completed
    - Item 1.3
        - Item 1.3.1
        - [x] Task 1.3.2 - Completed
- [x] Task 2 - Completed

As can be seen, there are three uncompleted tasks in total.

By using taskAny, it filters out the desired tasks with the desired structure. However, the task count is 4, instead of 3.

The code:

// Define function for expanding search to sub-tasks
const taskAny = function taskAny(t, f) {
  if (f(t)) return true;
  for (let sub of t.children) if (taskAny(sub, f)) return true;
  return false;
}


// Define function for hiding sub-tasks
const hideCompletedSubtasks = (t) => ({
	...t,
	children: t.children.filter(st => taskAny(st, st => (st.real && !st.checked) || st.text.includes("[["))).map(hideCompletedSubtasks)
})  // Show only uncompleted tasks and items that contain note links


// Search for tasks
const allTasks = dv.current()
	.file
	.tasks
	.where(t => taskAny(t, t => !t.checked))
	.sort(t => [t.due, t.p, t.i])


// Hide completed sub-tasks
allTasks.values = allTasks.values.map(hideCompletedSubtasks)


// Show tasks as a list
dv.header(5, "==Focus - Daily Actions==: " + allTasks.length + " tasks")

// dv.el("p", allTasks.length)  // Task count

for (let group of allTasks
	 .groupBy(t => t.p)
	 .sort(group => group.key, 'asc')) {
		dv.header(5, group.key)
		// dv.header(2, group.key.file.path)
		dv.taskList(group.rows
			.sort(t => [t.c, t.p, t.i])
			, true)
}

The result:

Without using taskAny, the filtered out tasks do not have the desired structure for sub-tasks but the task count is 3, which is correct.

The code:

// Define function for expanding search to sub-tasks
const taskAny = function taskAny(t, f) {
  if (f(t)) return true;
  for (let sub of t.children) if (taskAny(sub, f)) return true;
  return false;
}


// Define function for hiding sub-tasks
const hideCompletedSubtasks = (t) => ({
	...t,
	children: t.children.filter(st => (st.real && !st.checked) || st.text.includes("[[")).map(hideCompletedSubtasks)
})  // Show only uncompleted tasks and items that contain note links


// Search for tasks
const allTasks = dv.current()
	.file
	.tasks
	.where(t => !t.checked)
	.sort(t => [t.due, t.p, t.i])


// Hide completed sub-tasks
allTasks.values = allTasks.values.map(hideCompletedSubtasks)


// Show tasks as a list
dv.header(5, "==Focus - Daily Actions==: " + allTasks.length + " tasks")

// dv.el("p", allTasks.length)  // Task count

for (let group of allTasks
	 .groupBy(t => t.p)
	 .sort(group => group.key, 'asc')) {
		dv.header(5, group.key)
		// dv.header(2, group.key.file.path)
		dv.taskList(group.rows
			.sort(t => [t.c, t.p, t.i])
			, true)
}

The result:

[AnnaKornfeldSimpson]

There are 4 tasks rendered in the first version; although one is complete, it still counts in the number of tasks rendered. If you want that structure, you are inherently including some completed tasks. You could augment the count though, with your own calculation of only incomplete tasks from the group.rows and display it next to the header. Something like:

const incompleteInGroup = group.rows.filter(t => !t.checked).length().toString();
dv.header(5, (group.key + " " + incompleteInGroup + " incomplete"));

[wenlzhang]

@AnnaKornfeldSimpson Thanks for the help!

Now the result looks like the following, with the correct task count for uncompleted tasks:

The complete code is as follows:

// Define function for expanding search to sub-tasks
const taskAny = function taskAny(t, f) {
  if (f(t)) return true;
  for (let sub of t.children) if (taskAny(sub, f)) return true;
  return false;
}


// Define function for hiding sub-tasks
const hideCompletedSubtasks = (t) => ({
	...t,
	children: t.children.filter(st => taskAny(st, st => (st.real && !st.checked) || st.text.includes("[["))).map(hideCompletedSubtasks)
})  // Show only uncompleted tasks and items that contain note links


// Search for tasks
const allTasks = dv.current()
	.file
	.tasks
	.where(t => taskAny(t, t => !t.checked))
	.sort(t => [t.due, t.p, t.i])


// Hide completed sub-tasks
allTasks.values = allTasks.values.map(hideCompletedSubtasks)


// Show tasks as a list
dv.header(5, "==Focus - Daily Actions==: " + allTasks.length + " tasks")

// dv.el("p", allTasks.length)  // Task count

for (let group of allTasks
	 .groupBy(t => t.p)
	 .sort(group => group.key, 'asc')) {
		// dv.header(5, group.key)
		const incompleteInGroup = group.rows.filter(t => !t.checked).length.toString();
		dv.header(5, (group.key + " " + incompleteInGroup + " incomplete"));
		// dv.header(2, group.key.file.path)
		dv.taskList(group.rows
			.sort(t => [t.c, t.p, t.i])
			, true)
}

[AnnaKornfeldSimpson]

What is this t.p that you are grouping by? It doesn't seem to exist in your example. I was assuming it was a variable you had assigned somewhere. The header below with the count of 4 is coming from that very last true in your code, which tells taskList to group (within the groups you already have!) by file name. If you set that last true to false and do all grouping manually, you can put your count in the header and hide dataview's entirely. Good luck!

[wenlzhang]

What is this t.p that you are grouping by? It doesn't seem to exist in your example. I was assuming it was a variable you had assigned somewhere.

t.p is an inline field value denoting “priority” for tasks. For example, - [ ] Task [p::1] denotes a task with the priority 1. In the example I showed before, there are no such priority fields. That's why undefined was shown in the rendered results.

The header below with the count of 4 is coming from that very last true in your code, which tells taskList to group (within the groups you already have!) by file name. If you set that last true to false and do all grouping manually, you can put your count in the header and hide dataview's entirely. Good luck!

This works! Thanks for the hint!