#❓
Nested Searches:
SELECT FROM?
For example,
URL: Set yields nodes that contain a Url field.IS FIELD then extracts the pure field nodes, discarding their parent nodes.simply using and doesn't work: (URL==Set) AND (is:Field) . so nested searches needed.
==A:== Return just field values is a bit tricky
Path: >URL: (empty value)
IS FIELD will return field references, not field values
tried and yes, shouldn't use IS FIELD in this case.
Return just field values is a bit tricky
Path: >URL: (empty value)
exactly, I want to query the pure values of all the pasted links' URL fields as a single list.I've tried >Path: >URL: (empty value) , the first [https://tanacommunity.slack.com](https://tanacommunity.slack.com/) is as expected, but get some other results (not that pure):
raw links as pure texts like [https://tanacommunity.slack.com](https://tanacommunity.slack.com/) (whose father is a clickable link) as node refereces list is what i expect actually. (已编辑)
rather than clickable links (as directly pasted)
==A:== Add a line http or .com to the query
How can one open a new panel vertically (downward) rather than horizontally (to the right) in Tana?
==A:== Top / bottom panels have been deprecated
PKM/Apps/Tana 在 field object - field type - options - ==Sources of options== 中
通过 >:Path 只能 query build 出 schema > #Object of PKM app > Equivalents > Values (==这其实就是 Tana 的 Configure Field 页面顶部显示的路径==),
其中
- PARENT = Values
- GRANDPARENT = Equivalents
PKM/Apps/LogSeq 在 tag object - tag properties - property instance - ==(Default) value== 中,
通过 下述 当前 node 所有祖先(不包含自身)
{:query
[:find (pull ?a [*])
:in $ ?c %
:where
(ancestor ?c ?a)]
:inputs
[:current-block]
:rules
[(ancestor ?c ?p)
[?c :block/parent ?p]((ancestor ?c ?p)
[?c :block/parent ?p)
[(ancestor ?c ?a)
[?c :block/parent ?p]
(ancestor ?p ?a)]]}
只能 Query 出 property instance 一个值...
等价于 Tana 的 GRANDPARENT = Equivalents
==结论==:Tana 和 LogSeq 二者的 field templete 中的 query,在实例化后
没法 穿透 field 本身这个屏障,
以到达它 所属的 (Owned by in Tana) 实例化节点
I have two nodes, “itag” and “tag (base type)”, both tagged with #"Object of PKM app".
From a query builder node named “equivalents” placed under either node, it is easy to retrieve the other node sharing this tag, using query (Tagged:Object of PKM app) AND (NOT (Node name==PARENT) ) (where PARENT = “itag”).
However, within the “equivalents” field of #"Object of PKM app" itself, it is impossible to query any node with the same tag other than the node itself.
of no help EVEN using (Tagged:Object of PKM app) AND (NOT (Node name==GRANDPARENT) ) (where GRANDPARENT can only trace up to "equivalents" field node, rather than the GRANDGRANDPARENT “itag” node)
==A:== Search nodes within field configurations are unable to dynamically adapt to instances
pic1: when "equivalents" node as the child of the node "tag (base type)", all go smooth
pic2: when "equivalents" field as the field of the "tag (base type)", same query (as source of options) don't work
so I suppose we need GRANDGRANDPARENT to make things happen again.
==A:== It is not a GRAND^nPARENT issue, it is an instance adoption issue.
You can create a PARENT search node with a tag config template, then when you tag a node you can see the search within that instance adopts the PARENT of the instance.
Even within the tag template you can see the linter says INSTANCE to designate this will be filled once instantiated.This ability is not present with field config / field instance so the field dropdowns are not adaptable. Workaround is to do the former and put a search within the tag template.No timeline on when this will be addressed
Get it, is it something like below?
i.e. using a search node named "equivalents" in Tag Template to replace the "equivalents" field whose value is templated as a search node?

==A:== Yes your equivalents search is a perfect example of a search node that belongs to a tag template (#object of PKM app) and dynamically adapts to the PARENT of each instance
How can one obtain the instance node(without s) of a given tag (or field) within a tag’s field template (i.e., Sources of options)?
Owner node field provides value (though via a different mechanism)search node in tag template can access its PARENT node.I believe this missing "tunnel through field" feature should be introduced and is essential,
instance node's hierarchies but cannot use them as options,options yet cannot retrieve the node instance corresponding to that tag or field.From a development perspective, this may conflict with the tag’s “build title from fields” feature,
what we actually need is more than the name of the instance node, but also its hierarchical structures (like through PATH field),
which can be used from within field template and query as options.
Tunnel through field = Semantic Part Of + COMPONENTS REC https://tana.inc/faq#a-video-explanation-of-part-of-components-rec
Also you can write queries with PARENT.fieldname to insert the value of a field for a dynamic search
But yes as discussed in other thread, option-type field searches are blind to the instance details
Yes, I digged it out through your video series independently as well https://youtu.be/n1J0tZqb_6A?t=1354, and it almost lives in the meta graph.
Yet "PARENT.fieldname" lives in the instance graph?
(I found some similiar "${fieldname.(node's)fieldname}" in the same video https://www.youtube.com/watch?t=1274&v=n1J0tZqb_6A&feature=youtu.be, hah ;)
The issue appears to lie in Tana’s rigid distinction between the _meta graph_ and the instance graph? Certain scenarios require their integration. A comparable experiment in Logseq DB did not completely succeed but was partially effective -_-||.
Technically all the same graph. I think template / config path vs instance path might more accurately describe itWhen you build a PARENT search in a tag template / config it generates a new search node on each instance. This does not happen with searches in field config on instantiation
But yes I have personal use cases I would love to use this for as well if it existed
someday hopefully