Re: SQL/JSON features for v15

On 2022-09-01 Th 09:54, Nikita Glukhov wrote:
>
> On 31.08.2022 23:39, Nikita Glukhov wrote:
>
>> And here is a quick POC patch with an example for COPY and float4
> I decided to go further and use new API in SQL/JSON functions
> (even if it does not make real sense now).
>
> I have added function for checking expressions trees, special
> executor steps for handling errors in FuncExpr, CoerceViaIO,
> CoerceToDomain which are passed through ExprState.edata.
>
> Of course, there is still a lot of work:
> 1. JIT for new expression steps
> 2. Removal of subsidary ExprStates (needs another solution for
> ErrorData passing)
>   3. Checking of domain constraint expressions
> 4. Error handling in coercion to bytea
> 5. Error handling in json_populate_type()
> 6. Error handling in jsonb::type casts
> 7. ...
>
>
> Also I have added lazy creation of JSON_VALUE coercions, which was
> not present in previous patches. It really greatly speeds up JIT
> and reduces memory consumption. But it requires using of subsidary
> ExprStates.
>
>
> jsonb_sqljson test now fails because of points 4, 5, 6.

It looks like this needs to be rebased anyway.

I suggest just submitting the Input function stuff on its own, I think
that means not patches 3,4,15 at this stage. Maybe we would also need a
small test module to call the functions, or at least some of them.

The earlier we can get this in the earlier SQL/JSON patches based on it
can be considered.

A few comments:

. proissafe isn't really a very informative name. Safe for what? maybe
proerrorsafe or something would be better?

. I don't think we need the if test or else clause here:

+   if (edata)
+       return InputFunctionCallInternal(flinfo, str, typioparam,
typmod, edata);
+   else
+       return InputFunctionCall(flinfo, str, typioparam, typmod);

. I think we should probably cover float8 as well as float4, and there
might be some other odd gaps.

As mentioned previously, this should really go in a new thread, so
please don't reply to this but start a completely new thread.

cheers

andrew

--
Andrew Dunstan
EDB: https://www.enterprisedb.com