[Idea] Way to build timeline table and potential new function

[icytree]

I explored a little bit the idea of using Dataview to build timelines and it already works pretty well. For instance, assuming these files with that metadata :

file1:
- event:: 2021-01-02 Some info about this event.
- event:: 2019-02-03 Info about this one

file2:
- event:: 2020-05-04 Blah blah

file3:
- event:: 2005-09-25 Different info.
- event:: 2020-11-10 Details about this event.

It is possible to use a query like this :

```dataview
table event
from ""
flatten event
sort event asc
```

...to get all event in chronological order. The best thing about proceeding this way is that even if you only have a partial date (ex. only the year, as it can be possible I guess) it will still sort chronological (kind of).

Would it be possible to keep that great ability to sort complete and partial date (ex. YYYY-MM or YYYY) while having them not necessarily at the beginning ? For instance some event string could be like so : - event:: Information (2021-04) with extra details or like so - event:: Blah blah in 2022. and be combined with others with the format displayed earlier and the goal would be to parse the date from each of them to then flatten and sort chronologically.

I tried to workaround this using the current regexreplace but it is not so easy as I need to make it replace everything but the date and by doing so I am unable to make it differentiate between complete or partial date...

My questions would then be:

1. Is it be possible to implement a function that could pull out an ISO-8601 date format (even partial) from strings that contain more than only the date (like in my example above) ? I tried the date() function but it would not work unless the string only contained a date.
2. Would it be possible to implement a new function that could allows to extract part of a string based on regex pattern (some type of regexextract) ?

Any thoughts ?

[AB1908]

You can do most of this through DVJS. Writing custom sorts through DQL would be rather painful IMO.

[icytree]

I have no experience with JavaScript or DVJS, would you have some tips on how to go about pulling out dates of strings like I explained in my example ? Or basic pointers for someone in my situation ?

[AB1908]

Instead of DVJS, I wonder if we actually could get away with DQL.You can use regexmatch() to match parts of strings, pass the output to date() and then run a sort on it. So your thing might look like:

sort date(regexmatch("\d\d\d\d-\d\d-\d\d", field))

For trying out regex, use regexr.com. I've found it fairly helpful.

[icytree]

I think it can work really well with DQL. I tried yesterday with DVJS but got stuck trying to flatten the data. In DQL, the regexmatch() is limited as it only returns true or false depending if the pattern matches in the value. That is why I mentioned the addition of a new function for DQL that would allow to return the matching result of a regex pattern (kind of like the match() function in JS from what I could understand/test).

About how far I got with DQL, I simply decided to stick with a more rigid structure of entries atm to make it work better. The idea is to have these element in this precise order : the key (#log📝 ::), the timestamp (2022-04-16T10:09, 2022-04-16, 2022-04 or 2022) and an optional tag to indicate the type of log entry (#meeting🗣️, #idea💡, #issue🚩, ...). It can look something like this :

- #log📝 :: 2022-04-16T10:11 #howtoℹ️ format entries
- #log📝 :: 2022-04-16 #meeting🗣️ with this person about this [[Project]]

The query can look something like this :

```dataview
table 
	split(log📝, " ", 1) as timestamp, 
	regexreplace(log📝, "^(\d|-|\w|:)*", "") as entry-without-timestamp
from ""
flatten log📝 
sort log📝 asc
where log📝 
```

The date is separated from the rest using split() to get the 1st portion before the first space which allows better flexibility in the date/timestamp format. I also tried to extract the tags with limited success as it is a bit harder to match everything but the tags instead of trying to match any tags. Here's some of my attempts :

regexreplace(regexreplace(log📝, "^(\d|-|\w|:)*", ""), "[^#💡|ℹ️|🚩](\d|\s|\w)*$", "") as type, 	
regexreplace(regexreplace(log📝, "^(\d|-|\w|:)*", ""), "#\S*[^.*]", "") as entry

[AB1908]

This is quite the pain. It's a good FR as well. Would you do the honors?

[blacksmithgu]

I think it should not be too hard to add regexextract() which returns capture groups of regular expressions.

[jayambede]

I'm sorry to say this, but isn't that a question of how savvy you are with programming? I had a look into regex and imho this is rather complex and not easy to grasp, especially if you are not doing it professionally and your work focus is on something completely different.
For a creative writer, historian, journalist, a.s.o. creating a timeline from notes would be a really helpful feature, especially since the currently available plugins for these tasks are clunky and counter intuitive, therefore more or less a pain in the ass. That's why I'll file a detailed Feature Request for a TIMELINE Query type here