On 06/05/2015 04:48 PM, Tom Lane wrote:
> Andrew Dunstan <andrew@dunslane.net> writes:
>> Yeah, Good point. Actually, if my memory serves me correctly (always a
>> dubious bet), the avoidance of that kind of ambiguity is why we
>> introduced the #> and #>> operators in the first place, after going
>> round and round for a while on what the API would look like. I should
>> have remembered that when this came around. Mea culpa.
>> So probably the least invasive change would be to rename the text[]
>> variant operator to something like "#-" and rename the corresponding
>> function to jsonb_delete_path.
> Not sure that's a great choice of operator name; consider for example
> select 4#-1;
> It's not immediately obvious whether the "-" is meant as a separate
> unary minus. There are heuristics in the lexer that try to deal with
> cases like this, but it doesn't seem like a good plan to double down
> on such heuristics always doing the right thing.
>
>
Perhaps we should deprectae operator names ending in "-"?
cheers
andrew
