Is there an automated way to split a JSON(B) column into multiple columns in PostgreSQL?

Question

So I have a PostgreSQL (TimescaleDB) table that looks like this:

╔════════════════════════════╦═══════╦═══════╗
║            tags            ║ time  ║ value ║
╠════════════════════════════╬═══════╬═══════╣
║ {"a": "test", "c": "test"} ║ 10:24 ║   123 ║
║ {"b": "test", "c": "test"} ║ 10:25 ║   110 ║
║ {"b": "test"}              ║ 10:26 ║   130 ║
╚════════════════════════════╩═══════╩═══════╝

And I would like to split the JSON(B) column into multiple columns like this:

╔════════╦════════╦════════╦═══════╦═══════╗
║   a    ║   b    ║   c    ║ time  ║ value ║
╠════════╬════════╬════════╬═══════╬═══════╣
║ "test" ║        ║ "test" ║ 10:24 ║   123 ║
║        ║ "test" ║ "test" ║ 10:25 ║   110 ║
║        ║ "test" ║        ║ 10:26 ║   130 ║
╚════════╩════════╩════════╩═══════╩═══════╝

I looked into the JSON Processing Functions and it seems that those require finding all the JSON(B) attributes and types. Is there a way to do this automatically?

Answer 1

There is no way to make this dynamic. The number (and types) of all columns of a query must be known to the database when parsing the statement, long before it's actually executed.

---

If you always have the same structure you can create a type:

create type tag_type as (a text, b text, c text);

and then use jsonb_populate_record()

select (jsonb_populate_record(null::tag_type, tags)).*, time, value
from the_table;