Hi Bram,
On Sun, Sep 13, 2020 at 3:44 AM Bram Moolenaar <b...@moolenaar.net> wrote:
>
> > >> On Sat, Sep 12, 2020 at 1:09 PM Prabir Shrestha <
> > >> vim-dev-git...@256bit.org> wrote:
> > >>
> > >>> +1 for adding fuzzy match.
> > >>>
> > >>> A bit late to this. But few questions.
> > >>>
> > >>> How do we use matchfuzzy for a list of dictionary instead of item.
> > >>> Primary reason for this is LSP (language server protocol) makes
> heavy use
> > >>> of user_data. Can you add an example for sorting a list whose type
> is dict [
> > >>> { 'word': 'foo', user_data: 1}, { 'word': 'foobar', user_data: 2}, {
> > >>> 'word': 'bar', user_data: 3} ]. Should we have a 3rd optional
> parameter
> > >>> that takes a callback?
> > >>>
> > >>> let completeitems = [
> > >>> \ { 'word': 'foo', user_data: 1},
> > >>> \ { 'word': 'foobar', user_data: 2},
> > >>> \ { 'word': 'bar', user_data: 3}
> > >>> \ ]
> > >>> let filtertedlist = matchfuzzy(completeitems, 'foo',
> {item->item['word']})
> > >>>
> > >>>
> > >> That is a good idea. We can add a callback argument that will be
> called
> > >> for each item
> > >> if the list item is dictionary.
> > >>
> > >
> > > You can try the attached diff which implements the above and see how
> this
> > > works with a
> > > very large list of dictionaries.
> > >
> > >
> > I used the below functions to measure the time it takes to fuzzy search a
> > million entries
> > in a List and a Dict:
> >
> >
> > For a list, it took around a second to search all the entries and for a
> > dict,
> > it took around 5.5 seconds.
>
> A list is always going to be more efficient than a dict. It uses less
> memory and doesn't require hashing for the lookup.
>
> Also, when using a dict the key will be the same and rather meaningless?
>
The key is used to lookup the value in the dict which will be used for fuzzy
matching. So the value will be different for each list item.
> I can only imagine this being useful if you already have a list of
> dicts. How often does that happen?
>
>
For example, the getbufinfo(), gettagstack(), readdirex(), getqflist(),
getloclist(),
getmatches() and getwininfo() functions return a list of Dicts.
Without the support for fuzzy searching list of Dicts, the user needs
to first convert it to a List (while still retaining the rest of the
information
in the Dict) and then search.
For example, to fuzzy search 'eval' in a list of buffers returned by
getbufinfo():
With the support for fuzzy searching a list of Dicts using a 'key':
echo getbufinfo()->matchfuzzy('eval', 'name')
Generating a list of buffer names and then search (but the other
buffer related information is not kept):
echo getbufinfo()->map({_, v -> v.name})->matchfuzzy('quickfix.c')
Generating a list of lists with two elements and then searching using
the first element:
echo getbufinfo()->map({_, v -> [v.name, v]})->matchfuzzy('quickfix')
The first approach is more natural. In the last approach, the user needs
to generate a temporary list of lists for fuzzy search.
Regards,
Yegappan
> Using a list of lists, where the inner list has the text to fuzzy-match
> on as the first item and the rest of the list can be anything, would be
> the most efficient. Looking up the first item of a list is fast as
> well. Only creating this list of lists might take extra time, if you
> have an already existing structure.
>
> [['abcd', oneData], ['bcde', twoDta]]->matchfuzzy('cd')
>
>
>
--
--
You received this message from the "vim_dev" maillist.
Do not top-post! Type your reply below the text you are replying to.
For more information, visit http://www.vim.org/maillist.php
---
You received this message because you are subscribed to the Google Groups
"vim_dev" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to vim_dev+unsubscr...@googlegroups.com.
To view this discussion on the web visit
https://groups.google.com/d/msgid/vim_dev/CAAW7x7nbeAhUpG7_UUvVf0Z1YBAFx3COv3ZgnckMATaXZY584w%40mail.gmail.com.
