Re: Extension for PostgreSQL cast jsonb to hstore WIP - Mailing list pgsql-hackers

From ShadowGhost
Subject Re: Extension for PostgreSQL cast jsonb to hstore WIP
Date
Msg-id CAOsjGuOdsJcBPkmSyBkT9xrokgFGsTx_r6eEVPrXFgG_r+pTng@mail.gmail.com
Whole thread Raw
In response to Re: Extension for PostgreSQL cast jsonb to hstore WIP  (Andrew Dunstan <andrew@dunslane.net>)
Responses Re: Extension for PostgreSQL cast jsonb to hstore WIP
List pgsql-hackers
At the moment, this cast supports only these structures, as it was enough for my tasks
{str:numeric}
{str:str}
{str:bool}
{str:null}
But it's a great idea and I'll think about implementing it.

вт, 2 апр. 2024 г. в 19:48, Andrew Dunstan <andrew@dunslane.net>:

On 2024-04-02 Tu 07:07, ShadowGhost wrote:
> Hello all.
> Recently, when working with the hstore and json formats, I came across
> the fact that PostgreSQL has a cast of hstore to json, but there is no
> reverse cast. I thought it might make it more difficult to work with
> these formats. And I decided to make a cast json in the hstore. I used
> the built-in jsonb structure to create it and may have introduced
> methods to increase efficiency by 25% than converting the form
> jsonb->text->hstore. Which of course is a good fact. I also wrote
> regression tests to check the performance. I think this extension will
> improve the work with jsonb and hstore in PostgreSQL.
> If you've read this far, thank you for your interest, and I hope you
> enjoy this extension!
>

One reason we don't have such a cast is that hstore has a flat
structure, while json is tree structured, and it's not always an object
/ hash. Thus it's easy to reliably cast hstore to json but far less easy
to cast json to hstore in the general case.

What do you propose to do in the case or json consisting of scalars, or
arrays, or with nested elements?


cheers


andrew



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

pgsql-hackers by date:

Previous
From: Kartyshov Ivan
Date:
Subject: Re: [HACKERS] make async slave to wait for lsn to be replayed
Next
From: Tom Lane
Date:
Subject: Re: Detoasting optionally to make Explain-Analyze less misleading