On Sep 10, 2025, at 07:53, Alexandra Wang <alexandra.wang.oss@gmail.com> wrote:
I think this (jb)[0] case is rather trivial. However, since it's been
behaving the pre-standard way since PG14, I hesitate to change the
existing behavior as part of my patches, and I feel it could be a
bikeshedding kind of discussion in a separate thread.
I am fine to retain the PG14 behavior. But the confusion part is:
```
evantest=# select ('{"name": "Alice", "birthday": {"year": 2000, "month": 8}}'::jsonb)[0]['birthday']['year'];
jsonb
-------
(1 row)
evantest=# select ('{"name": "Alice", "birthday": {"year": 2000, "month": 8}}'::jsonb)[0]['birthday'].year;
year
------
2000
(1 row)
```
As long as the expression containing a dot notation, “[0]” behaves differently from PG14.
I agree we should not recommend the mixed usages of `[`a`]` and `.a`, but we haven’t prevented from that. Given they are different meanings, `[‘a’]` implies strict mode and `.a` implies lax mode, for a complex JSON object, users may intentionally mix use both, which sounds reasonable.
I don’t see you have updated any documentation yet. I don’t want to block you because of this issue. As long as you state the behavior clearly in the document, I am happy to let it go.
Best regards,
--
Chao Li (Evan)
HighGo Software Co., Ltd.
https://www.highgo.com/
