Totally new to Logseq (< 1 month), have some experience with Omnioutliner but that it. This video was useful but covered the wrong material for me, once the syntax became apparent the AND, OR ,NOT, BETWEEN queries based on links & properties are reasonaby simple; the issue I have is understanding what metadata is intrinsically available (for example how do I know 'todo' is a thing in the context of a query) and how to create metadata in a way that is useful for querying. For example my collection of recipies lives, today, in a linux folder structure; library/{beef,chicken...} library/{Indian,Persian...} and similar classifications strung together with symbolic links and a python script that create a file:// based website by parsing the file-system. Clearly my directories beef etc. can map to properties or pages. In terms of querying the recipe collection, what i'm lacking is the knowledge to decide if a a recipe should be a page or a block; where the tags should go to be useful for searching, how to query all the tags i've previously defined so as to minimize duplication. Part of this is a paradigm shit issue but the current documentation makes it difficult to bridge that paradigm gap. I'm finding that videos are either too trivial (no offence but this video falls in that category from my perspective, it sortof stops at the point I need to start) or superficial in the sense they show how to do X but don't provide the background information/context necessary to understand how to apply the same principles to other situations. What I need is a video that spends at most 5ish minutes going over the simple filter syntax then dives into using 'intrinsic' metadata to create query result pages and how to structure you own metadata to best aid queries. I've spent 40+ years in IT Develpment but find myself in the situation where I need the "Janet & John guide to building a nuclear bomb" when it comes to Logseq. If you can point me at suitable videos it would be appreciated.
In my day they were called prompts and they were undeceptively simple yet powerful. C:\dir in Windows or ~ls in Linux meant: file=$s where * $s. were usually a list of 3, 16 colour 20x20 lores buttmaps that had the power to summon a whole van of parties.
There was a question at the end "is it possible to sort the results of a query". If you use the table view, you can click on a column to sort the results based on that columns value.
I really like Logseq (and donate to the development fund), but it needs a much simpler query capability! I vote for the visual query builder! I am confused about what queries can find. You say they can search for any term, but it appears they can only search for linked references. Can queries find any text? Or just linked references?
This is all mentioned in the session, including use cases. Avoid queries for simple search operations, especially as a beginner. Queries can look for text only, but that makes them much less precise. Hence the recommendation to use search terms that you’ve created more intentionally. We agree that writing queries can be made simpler with a UI, which we will further in investigate. If you have suggestions, please let us know on the forum: discuss.logseq.com/c/feature-requests/7
For me it is missing how I plan it. The concept behind. In which cases is this useful. Isn't it a lot of work if you add for every block a property. It is good to see how it works with mood and sad etc. But if I think I have to add for every block all possible properties then it very soon becomes very complex. So what is the strategy?
Don't overthink structure too much, especially not in the beginning when you're new to Logseq. First get used to the block-based way of thinking before creating complex queries with properties. Once you're ready, you'll see that templates are the best way to create consistency in how you use properties. Adding them manually each time is error-prone.
Super informative, thanks. Any ideas about how we could expand the `between` query you used to list journal entries between given dates to also include dated entries from other pages? I keep track of projects in their own pages. Each day I want to jot down some notes on that project, I create a new parent node in that page that's simply the day's date. The children to this node are all the notes for that day. But these don't show up when I perform a {{query (between [[DATE A]] [[DATE B]])}} since that only pulls in journal entries.
We're planning to expand the queries to work more with the underlying metadata, but for now you could use a property like created-at or modified-at: docs.logseq.com/#/page/62559bea-6b63-4536-bb6e-bd5aa6420043
31:50 I was sad when I realised the chronological order of your times and your days are inverse. Noob conjure mistake. Upside down is the new downside up am I right? lol ;D Im kidding it is great....ly disorientating... lol sorry cant help myself
Totally new to Logseq (< 1 month), have some experience with Omnioutliner but that it. This video was useful but covered the wrong material for me, once the syntax became apparent the AND, OR ,NOT, BETWEEN queries based on links & properties are reasonaby simple; the issue I have is understanding what metadata is intrinsically available (for example how do I know 'todo' is a thing in the context of a query) and how to create metadata in a way that is useful for querying. For example my collection of recipies lives, today, in a linux folder structure; library/{beef,chicken...} library/{Indian,Persian...} and similar classifications strung together with symbolic links and a python script that create a file:// based website by parsing the file-system.
Clearly my directories beef etc. can map to properties or pages. In terms of querying the recipe collection, what i'm lacking is the knowledge to decide if a a recipe should be a page or a block; where the tags should go to be useful for searching, how to query all the tags i've previously defined so as to minimize duplication. Part of this is a paradigm shit issue but the current documentation makes it difficult to bridge that paradigm gap. I'm finding that videos are either too trivial (no offence but this video falls in that category from my perspective, it sortof stops at the point I need to start) or superficial in the sense they show how to do X but don't provide the background information/context necessary to understand how to apply the same principles to other situations.
What I need is a video that spends at most 5ish minutes going over the simple filter syntax then dives into using 'intrinsic' metadata to create query result pages and how to structure you own metadata to best aid queries. I've spent 40+ years in IT Develpment but find myself in the situation where I need the "Janet & John guide to building a nuclear bomb" when it comes to Logseq. If you can point me at suitable videos it would be appreciated.
In my day they were called prompts and they were undeceptively simple yet powerful. C:\dir in Windows or ~ls in Linux meant: file=$s where * $s. were usually a list of 3, 16 colour 20x20 lores buttmaps that had the power to summon a whole van of parties.
There was a question at the end "is it possible to sort the results of a query". If you use the table view, you can click on a column to sort the results based on that columns value.
That option indeed exists, but only works in combination with properties. For queries that don't use properties the sort-by filter exists.
Thank you! I've only recently moved from Roam to Logseq and these sessions are really useful.
Have you started producing really long logs yet?
I really like Logseq (and donate to the development fund), but it needs a much simpler query capability! I vote for the visual query builder!
I am confused about what queries can find. You say they can search for any term, but it appears they can only search for linked references. Can queries find any text? Or just linked references?
This is all mentioned in the session, including use cases.
Avoid queries for simple search operations, especially as a beginner. Queries can look for text only, but that makes them much less precise. Hence the recommendation to use search terms that you’ve created more intentionally.
We agree that writing queries can be made simpler with a UI, which we will further in investigate. If you have suggestions, please let us know on the forum: discuss.logseq.com/c/feature-requests/7
For me it is missing how I plan it. The concept behind. In which cases is this useful. Isn't it a lot of work if you add for every block a property. It is good to see how it works with mood and sad etc. But if I think I have to add for every block all possible properties then it very soon becomes very complex. So what is the strategy?
Don't overthink structure too much, especially not in the beginning when you're new to Logseq. First get used to the block-based way of thinking before creating complex queries with properties.
Once you're ready, you'll see that templates are the best way to create consistency in how you use properties. Adding them manually each time is error-prone.
Thanks for sharing! Very useful
Thank you, very helpful for me!
Excellent session! Keep it up 👏
Thanks!
Got some cool ideas from your session.
Looking forward for new videos 🙌
Another awesome 👏 session! Thank you!
Super informative, thanks. Any ideas about how we could expand the `between` query you used to list journal entries between given dates to also include dated entries from other pages? I keep track of projects in their own pages. Each day I want to jot down some notes on that project, I create a new parent node in that page that's simply the day's date. The children to this node are all the notes for that day. But these don't show up when I perform a
{{query (between [[DATE A]] [[DATE B]])}}
since that only pulls in journal entries.
We're planning to expand the queries to work more with the underlying metadata, but for now you could use a property like created-at or modified-at: docs.logseq.com/#/page/62559bea-6b63-4536-bb6e-bd5aa6420043
@@Logseq Ah, thank you! The answer is always reading the docs more carefully...
Buen video, muy bien explicado para entrar en el concepto de Query 🙌
The missing link and what ought to be in the todo is to get this onto the Playstore, but first you would have to iron out the compatibility issues.
31:50 I was sad when I realised the chronological order of your times and your days are inverse. Noob conjure mistake. Upside down is the new downside up am I right? lol ;D Im kidding it is great....ly disorientating... lol sorry cant help myself
You okay Ramses? You use "sad" as an example concerningly often 😅
Nothing to worry about, just an example! 😂