Fix #14842 #14885
Fix #14842 #14885
Conversation
|
ya, that looks cool. I'm up for adding something like that. I'm just not sure how we'd indicate the default row/record. 🤔 Another thing I've wanted to do for a long time is specify column widths manually in a record. something like |
If adding a But performance will had to be somewhat better. In this regard I wanna highlight an approach which some tools take, in regard of large datasets. It's considered also a performance optimization. Hmmmm now thinking about .... 😄 I guess even current approach could be optimized. Well it's a small but if there's lots of columns measurable optimization. |
|
We're definitely up for buffering x amount to estimate widths. I wrote something like this with our original table implementation a long time ago, but it was really to try and get emoji to line up properly. I now know that emoji is a hard problem, so I'm not as concerned with it these days. I'd still like to be able to manually specify column widths at some point because the estimates will never be 100% perfect. |
fdncred
added
tabled
|
Thanks! This PR looks good. I'm looking forward to seeing the other parts of our conversation prototyped somewhere. |
Sorry was a little bit busy
close #14842
I've added a test but I'd check if it solved it.
cc: @fdncred
Unrelated
Recently got a pretty good format idea (zhiburt/tabled#472)
Just wanna highlight that we could probably experiment with it, if it being a bit elaborated.
It's sort of KV table which nushell already has,
But it's more for a default table where each row/record being rendered as a KV table.
It's not something super nice I guess but maybe it could get some appliance.
So yes pointing it out just in case.
Like these.
PS: Now thinking about it,
it's sort of like doing a iteration over rows and building a current KV table,
Which is interesting cause we could do it row by row, in which case doing CTRLC would not ruin build but got some data rendered.
All though it's a different kind of approach. Just saying.