In a previous HN comment [1] I showed how you can leverage PRQL to make your RDBM's JSON functionality more ergonomic. For example
```sh
> prqlc compile <<EOF - | duckdb
let get = path obj -> s"""{obj} -> {path}"""
let getstr = path obj -> s"""{obj} ->> {path}"""
let extract = obj path -> s"""json_extract({obj}, {path})"""
from [{data='{"duck": [1, 2, 3]}'}]
select { (data | get '$.duck[0]'), (data | getstr '$.duck[1]'), extract data '$.duck[2]'}
EOF
┌───────────────────────┬──────────────────────────┬───────────────────────────────────┐
│ "data" -> '$.duck[0]' │ ("data" ->> '$.duck[1]') │ json_extract("data", '$.duck[2]') │
│ json │ varchar │ json │
├───────────────────────┼──────────────────────────┼───────────────────────────────────┤
│ 1 │ 2 │ 3 │
└───────────────────────┴──────────────────────────┴───────────────────────────────────┘
```
Unfortunately I don't think this really addresses the grandparent comment though because you're still using jsonpath type expressions to unpack the JSON objects. If you really wanted to use PRQL for everything you would have to first convert and flatten your JSON data into relational tables.
Unfortunately I haven't had time to maintain it so it is now archived and out of date. I hope that I might get a chance to update it again. More has happened since then and there are low hanging fruit to make it more usable, for example adding connector_arrow (https://github.com/aljazerzen/connector_arrow) support for other databases, etc...
Quick example of how things looked with pq:
```sh
$ pq --from i=invoices.csv "from i | take 5"
+------------+-------------+-------------------------------+-------------------------+--------------+---------------+-----------------+---------------------+-------+
| invoice_id | customer_id | invoice_date | billing_address | billing_city | billing_state | billing_country | billing_postal_code | total |
+------------+-------------+-------------------------------+-------------------------+--------------+---------------+-----------------+---------------------+-------+
| 1 | 2 | 2009-01-01T00:00:00.000000000 | Theodor-Heuss-Straße 34 | Stuttgart | | Germany | 70174 | 1.98 |
| 2 | 4 | 2009-01-02T00:00:00.000000000 | Ullevålsveien 14 | Oslo | | Norway | 0171 | 3.96 |
| 3 | 8 | 2009-01-03T00:00:00.000000000 | Grétrystraat 63 | Brussels | | Belgium | 1000 | 5.94 |
| 4 | 14 | 2009-01-06T00:00:00.000000000 | 8210 111 ST NW | Edmonton | AB | Canada | T6G 2C7 | 8.91 |
| 5 | 23 | 2009-01-11T00:00:00.000000000 | 69 Salem Street | Boston | MA | USA | 2113 | 13.86 |
+------------+-------------+-------------------------------+-------------------------+--------------+---------------+-----------------+---------------------+-------+
$ # When there is only one input table then this automatically becomes the source relation, i.e. `from i | ` is prepended to the query
$ # so this can be simplified to:
$ pq --from invoices.csv "take 5"
...
```
[1] https://prql-lang.org/