How to find all tasks under blocks tagged with project namespaces?

[Michael_Rienitz]

Hi there, i am new to advanced queries in Logseq.
I found your hint for queries in namespaces and tried to apply it to my own example, but get no results.

This is my example. I try to get a query that shows the two tasks in the sub-pages of the namespace.

This is my query. I only changed the String “TODO” to “LATER”.

Can you give me an advice, what I am missing?
Thanks in advance.

[Siferiax]

This goes from relatively simple to relatively complex pretty fast.
First of :block/refs doesn’t get what you need.
Also you want the block on the page and not the parent of the block, though in this case that is the same, but if your task is under a parent block you won’t retrieve it.
Next getting a namespace is only 1 level deep, and here’s where the complexity comes from.
I’m going to assume you don’t want that to be static
So I use rules to create a recursive search to find all pages under project, no matter their dept.
Here’s the query I created:

#+BEGIN_QUERY
{:title "Find Sub-Tasks"
 :query [:find (pull ?b [*])
   :in $ %
   :where
     [?ns :block/name "project"]
     (check-ns ?ns ?p)
     [?b :block/page ?p]
     [?b :block/marker ?marker]
     [(contains? #{"LATER"} ?marker)]
 ]
 :rules [
   [(check-ns ?ns ?page)
     [?page :block/namespace ?ns]
   ]
   [(check-ns ?ns ?page)
     [?page :block/namespace ?t]
     (check-ns ?ns ?t)
   ]
 ]
}
#+END_QUERY

[Michael_Rienitz]

Perfect! It works!
Now I really can draw the Power of Logseq for my purposes.
Thanks, mighty wizard.
I do not understand the magic yet, but can use it and am eager to learn more.

[Michael_Rienitz]

When i put “project” (or the final target string) in lowercase letters, it works as intentioned.
Next i’ll try to filter on tasks with priority “A”.
Maybe i can do it without help from a magician.

[Michael_Rienitz]

Dear Siferiax, I do not want to be greedy but could you extend your helpful Query so that it Filters Namespace + Tasks + Priority “A”
As a beginner i still struggle a little with learning datalog.

[Siferiax]

Yes. :block/name is always lowercase. I forgot to mention, I generally add that for clarity!
Pages have 2 names, one is lowercase and stored in the attribute :block/name and the other is their original name case stored in :block/original-name.
I generally use :block/name so we don’t have to think how exactly we wrote the name of a page

[Siferiax]

For filtering priority we have the attribute :block/priority
So we will need to add the line: [?b :block/priority "A"] to the query.
Alternatively we can do this the same way as we did for the task itself.

[?b :block/priority ?prio]
[(contains? #{"A" "B"} ?prio)]

In this case finding both A and B priority tasks. Just to illustrate how you can use multiple values for the contains? function.

Complete query for clarity:

#+BEGIN_QUERY
{:title "Find Sub-Tasks"
 :query [:find (pull ?b [*])
   :in $ %
   :where
     [?ns :block/name "project"]
     (check-ns ?ns ?p)
     [?b :block/page ?p]
     [?b :block/marker ?marker]
     [(contains? #{"LATER"} ?marker)]
     [?b :block/priority ?prio]
     [(contains? #{"A"} ?prio)]
 ]
 :rules [
   [(check-ns ?ns ?page)
     [?page :block/namespace ?ns]
   ]
   [(check-ns ?ns ?page)
     [?page :block/namespace ?t]
     (check-ns ?ns ?t)
   ]
 ]
}
#+END_QUERY

[Michael_Rienitz]

Hi there @Siferiax , many thanks for the solution… its nearly perfect.
If you allow one last question:
Sometimes i would also note some tasks in the root layer of a namespace (“project/…”).
Is it possible to include these tasks with priority into the query results?

Best regards, already a fan…

[logseq-lion]

Someone should develop a GPT based AI that can transform natural language queries into advanced logseq queries! I wonder if anyone is thinking about this?

[Michael_Rienitz]

Hi there,
i already found the great Query-Builder-Tool from adxsoft

Query Building Tools

Two tools for building test graphs and advanced queries GitHub - adxsoft/buildlogseqtestgraph: Build a test logseq graph with generated data and query testing capabilities - python app includes a generated test graph with query testing (only tested on Mac OS so far) GitHub - adxsoft/logseqadvancedquerybuilder - online pyscript app that simplifies building advanced queries (Seems to work ok on Desktop and Mobile) I have used these for my testing purposes so they may help others especially thos…

but couldn’t figure out how to build a recursive query for the namespaces like the one @Siferiax provided. Maybe its just a matter of time learning the nuances of datalog.
For now the hints from Siferiax provides a very powerful solution for project management in Logseq.

[Michael_Rienitz]

There seems to be another question/solution that tackles the problem of queries in hierarchies.

Query todos recursively under ProjectA/Module1/Sub-Module1

Hi all, I have the following page structures to track my projects ProjectA/Module1/Sub-Module1 ProjectA/Module1/Sub-Module2 ProjectA/Module2/Sub-Module1 I would like a list of all todos in Module1 - i.e. todos in Module1, Sub-Module1, Sub-Module2 - to show in Module1 for a query. Similarly for ProjectA all Todos recursively under the pages reference in it. Can someone please suggest a query to add in each of the pages, such that it shows the todos of the pages referenced under it? Regard…

[Siferiax]

Yes, definitely!
Change [?b :block/page ?p] to:

(or
  [?b :block/page ?ns]
  [?b :block/page ?p]
)

[Siferiax]

Yeah that solution does basically the same thing as my solution

[Michael_Rienitz]

Thanks mighty Query-Wizard!

[logseq-lion]

So this query you suggested works

#+BEGIN_QUERY
{:title "🔨 All projects"
  :query [:find (pull ?b [*])
          :where
          [?p :block/name "projects"]
          [?par :block/refs ?p]
          (or [?b :block/parent ?par]
     [?b :block/refs ?p])
          [?b :block/marker ?marker]
          [(contains? #{"TODO"} ?marker)]
  ]
:collapsed? true
}
#+END_QUERY

But now I’m trying the following as the inverse - i.e. find all todos that don’t have any project tags, but it’s finding no results (despite the fact that I’ve created some test todos without project tags). Any idea how I can fix this? Thanks very much in advance!

#+BEGIN_QUERY
{:title "🔨 Inbox"
  :query [:find (pull ?b [*])
          :where
          [?p :block/name "projects"]
          [?par :block/refs ?p]
          (not (or [?b :block/parent ?par]
     [?b :block/refs ?p]))
          [?b :block/marker ?marker]
          [(contains? #{"TODO"} ?marker)]
  ]
:collapsed? false
}
#+END_QUERY

[Siferiax]

Check out this post:

Queries for task management

A fun one today! Tasks that don’t reference a page regardless of nesting This will exclude any tasks that are referencing a page or that one of the parent blocks references a page. This query uses rules for recursive searching. To change this to reference instead of don’t reference, just remove the (not and closing ) around (check-ref ?p ?b) #+BEGIN_QUERY {:title " TODO" :query [:find (pull ?b [*]) :in $ % ; % is used to pull in the rules below. :where [?b :block/marker ?marker]…

(Sorry doing this from a hospital bed on my phone but thought I could at least point you in the right direction)

[logseq-lion]

Omg, you’re my savior - this works perfectly. For whatever reason you’re in the hospital - I hope you get well soon!

[Siferiax]

Thanks. Had emergency surgery, but I’m back home and recovering just fine.

[veliger]

This thread is so helpful! I would like to sort the output by project - right now it is sorted by date of journal entry.

I tried adding:

:result-transform (fn [result]
                                 (sort-by (fn [h]
                                                  (get h :block/name)) result))

With no change

[Siferiax]

It outputs blocks, not pages. So we need to first get the page.
(get h :block/name)
Needs to be (get-in h [:block/page :block/name])