Why do these searches (taken individually) which I thought would match with everything, in fact match with nothing
How should they be rewritten to match with everything ?
It’s still not clear despite reading the appendix twice. I must not be very smart.
name:~*
name:*
text:*
thanks in advance for your time and help
If you look at the expanded version of this search, you might be able to spot it yourself:
What you’re asking for is possible only in quantum physics, where a cat can be alive and dead at the same time. In real life, you’ll not find a name that is anything and nothing simultaneously.
thank you @chrillek and @BLUEFROG
Please keep in mind that I am trying to understand search operators.
My questions is not for a specific search. It’s for 2 reasons:because I am writing a Keyboard Maestro macro to handle searches and also because I would like to understand the operator syntax.
I would like to know which operator to use to match any name and any content (2 separate questions).
I read the appendix 2 times taking notes and I will don’t understand certain things.
What is logically wrong with no matches with
I don’t understand what you’re asking here. If you want to find a name that matches “anything”, simply do not search for a name. You’re not interested in it anyway, so there’s no need to specify a search condition for it. The logical flaw in this attempt is that there is no “I don’t care, but I’m asking anyway” operator. If there were, you could use it.
It would still make pitifully little sense.
Same thing. You want everything. So why even search for it.
It’s like telling your favorite online book store “Hey, I want a book with any title and any ISBN”. What do you expect to get? A single book? Which one?
because I am writing a Keyboard Maestro macro in which case it makes sense. I need to enter every predicate into the macro as long as they match all items.
All is working fine such as kind:any extension==any and others, but I am missing content (text:) and name
From your point of view, I agree, it does not make sense.
thank you for your reply
From the example in the Help: Name contains tech → name:~tech.
And a thorough explanation of the two…
For an item named DEVONtechnologies:
name:~tech matches as the substring is contained anywhere in the name.
name:tech does not match as there is no word tech in the name.
name:tech* does not match as there is no word beginning with tech.
name:*tech does not match as the substring is found but not at the end of the word.
name:*tech* matches as tech is matched with text before and after it. However, unless you have a specific purpose, using the contains operator is more succinct.
PS: Bear in mind, the wildcard is not RegEx.
Why? If you’re looking for “creation date before 2015”, what’s the point in adding “type: any”? That just gets thrown out anyway by DT before it starts searching (hopefully).
OK all clear now. I will find a workaround. thanks very much
Because I am working in the opposite direction using Keyboard Maestro.
I populate the search pane with all matches and all filters (when possible) and as a second step gradually trim down the list of matches by interacting with the list of filters as in (2).
I understand your point, and please understand the problems that I am attempting to solve. DevonThink is complex full of data, and fast and efficient searches are essential. I also try to minimize the number of mouse clicks.
. if I enter searches in the search pane, I have to remember a plethora of operators which in my case means that I have to have a printout of the appendix under my nose
. If I take the other route, which is to manipulate the list of filters (2 in the ScreenShot below), it becomes tedious to navigate the drop down list of each filter.
thanks again
You’re replacing the complexity of one app with complexity in another app. Adding friction. Basically, you’d have to duplicate DT’s search interface one way or another in KM. Either you have everything that exists in DT in KM, then why would that be “easier” to use (and in what sense?) or it’s incomplete.
In my opinion, you are overengineering it. No one can remember all the search options. Nor all HTML elements with their attributes, nor all CSS selectors with their subtleties. There are things that you need most, and you’ll remember those. For the rest, you can look them up. And if you happen to use one or more of the looked-up operators/prefixes/whatever, you’ll remember them, too.
You have a good point that I will keep in mind.
This is what it looks like (just starting, not complete and not in the right order, and hotkey triggers for each menu item).
A Keyboard Maestro generated menu based search triggered from anywhere inside and out of DevonThink (if out of DevonThink, will activate DevonThink first). Extremely fast. IMO faster than typing in the search pane with the manual in front of me, or navigating through filters. Obviously not all options are there , but the end result will cover 90% of my searches.
That’s what I imagined. A functional duplicate of the original, with tons of shortcuts to remember. If you don’t remember them, you have to look them up…
Or use the advanced search without having to worry about the search syntax.
it’s an interactive palette, always displayed until closed
I think what rufus123 means is that you don’t have to remember anything. The menu items are triggered by clicking. If you want a shortcut, it is displayed in the same palette to the right of the name. 
I got that.
And that is exactly what DT’s search dialog offers: Clicking on the terms and conditions. But that one is complete and offers a reasonable UX experience. If someone wants to re-write that in KM, why not. But I can’t see the advantage over what is already offered in DT. Apart from improving the personal KM skill set.
So why do you write
Only rufus123 can tell you the rest.
and hotkey triggers for each menu item
