Hello, I am using Logseq to (among other things) process academic papers in my reading list. The way I mark the process is by creating a block in the paper page with a TODO marker and the text “Preparation” and I have under it two child blocks [[Highlighting]] and [[Note structuring]].
I want to create a query to find all papers that are e.g. done with the Highlighting but the Note structuring is “TODO” or “DOING”. I am very new to Datalog and I tried to create something on my own, but without any success. Any help would be much appreciated.
1 Like
Something like this:
#+BEGIN_QUERY
{:inputs [:today]
:query [:find (pull ?b [*])
:in $ ?today
:where
[?highlighting :block/original-name "Highlighting"]
[?structuring :block/original-name "Note structuring"]
[?c1 :block/refs ?highlighting]
[?c1 :block/marker ?marker1]
[(contains? #{"DONE"} ?marker1)]
[?c1 :block/parent ?b]
[?c2 :block/parent ?b]
[?c2 :block/refs ?structuring]
[?c2 :block/marker ?marker2]
[(contains? #{"TODO" or "DOING"} ?marker2)]
]
}
#+END_QUERY
I have a quite similar issue where I have a Journal with a section called “Tasks” and one called “Notes”.
Example:
### **Tasks**
- TODO Task 1
- TODO Task 2
### **Notes**
- Note 1
- Note 2
- TODO Something to do
- Note 3
- TODO Another thing to do
I need a query which only returns the TODOs nested under the “Notes” section but not the ones under “Tasks”. No matter what I try, I either get all TODOs from all Journals or nothing at all.
I tried to rewrite @mentaloid’s example as follows but get nothing in return:
#+BEGIN_QUERY
{ :inputs [:today]
:query [:find (pull ?b [*])
:in $ ?today
:where
[?notes :block/original-name "Notes"]
[?c2 :block/parent ?b]
[?c2 :block/refs ?notes]
[?c2 :block/marker ?marker2]
[(contains? #{"TODO" "LATER"} ?marker2)]
]
}
#+END_QUERY
I also tried the example from Queries for task management - Customization / Look what I built - Logseq and changed it to show all from today, which works. But now I also have the tasks from the “Tasks” section, which I dont want.
Now I am out of ideas. Anybody who can guide me in the right direction? 
- Your tasks are nested deeper than one level from their section’s root.
- That may require recursive lookup (read lower).
- Could still be avoided if the number of levels under the root is always the same.
- That may require recursive lookup (read lower).
- To filter by section, you need a way to identify the sections. That can be done by either:
- using content search for
"**Notes**"- slower, because of string search
- needs recursive lookup, unless
**Notes**is at known level - requires no references
- using a reference, either:
- in the direct parent or the task itself, e.g.
#note- avoids recursive lookup
- requires more references (one per note or task)
- in the section title, e.g.
**[[Notes]]**- needs recursive lookup, unless the reference is at known level
- requires less references (one per section)
- can also facilitate a fast query of all such sections
- in the direct parent or the task itself, e.g.
- using content search for
- Which one of the above approaches do you prefer?
- My personal suggestion is for the last approach, but you may differ.
Ha! Didn’t think I get an answer that fast so I did a bit of back and forth with ChatGPT on that topic. It now works for me with the following query:
#+BEGIN_QUERY
{
:title [:h4 [:b "From today's notes"]]
:query [:find (pull ?task [*])
:in $ ?day
:where
[?p :block/journal-day ?day] ;; Find today's journal page
[?section :block/parent ?p] ;; Get the parent of the section (the journal page)
[?section :block/content ?section-content]
[(clojure.string/includes? ?section-content "Notes")] ;; Ensure the section is "Notes"
[?note :block/parent ?section] ;; Find blocks under the "Notes" section
[?task :block/parent ?note] ;; Check for tasks under those blocks
[?task :block/marker ?m] ;; Filter for TODO tasks
(not [(contains? #{"CANCELED"} ?m)])
]
:inputs [:today]
:table-view? false
:breadcrumb-show? false
:result-transform (fn [n] (map (fn [x] (ident x)) n))
}
#+END_QUERY
Thanks anyways for your help.