út 23. 1. 2024 v 11:38 odesílatel Christoph Berg <myon@debian.org> napsal:
Re: Pavel Stehule > It looks great for simple queries, but if somebody uses it like SELECT * > FROM pg_proc \gedit
What's wrong with that? If the pager can handle the amount of data, the editor can do that as well. (If not, the fix is to just not run the command, and not blame the feature.)
just editing wide or long command or extra long strings can be unfriendly
> I almost sure so \gedit is wrong name for this feature.
I'm open for suggestions.
I have not too many ideas. The problem is in missing relation between "edit" and "update and execute"
> Can be nice if we are able: > > a) export data set in some readable format > > b) be possible to use more command in pipes > > some like > > select start, call, qrg, name from log where cty = 'CE9' order by start > \gpipexec(tsv) mypipe | bash update_pattern.sh > tmpfile; vi tmpfile; cat > tmpfile > mypipe
Well yeah, that's still a lot of typing.
it should not be problem. You can hide long strings to psql variables
> I understand your motivation well, but I don't like your proposal because > too many different things are pushed to one feature, and it is designed for > a single purpose.
It's one feature for one purpose. And the patch isn't *that* huge. Did I make the mistake of adding documentation, extra command options, and tab completion in v1?
no - I have problem so one command does editing, generating write SQL commands (updates) and their execution