issue_comments

FWIW I implemented something similar with a bit of plugin code:

```python @hookimpl def canned_queries(datasette: Datasette, database: str) -> Mapping[str, str]: # load "canned queries" from the filesystem under # www/sql/db/query_name.sql queries = {}

sqldir = Path(__file__).parent.parent / "sql"
if database:
    sqldir = sqldir / database

if not sqldir.is_dir():
    return queries

for f in sqldir.glob('*.sql'):
    try:
        sql = f.read_text('utf8').strip()
        if not len(sql):
            log(f"Skipping empty canned query file: {f}")
            continue
        queries[f.stem] = { "sql": sql }
    except OSError as err:
        log(err)

return queries

```

I actually think it would be a useful thing to add support for in datasette. It wouldn't be difficult to unwind an array of params and add the placeholders automatically.

The easiest or most straightforward thing to do is to use named parameters like:

sql select * where key IN (:p1, :p2, :p3)

And simply construct the list of placeholders dynamically based on the number of values.

Doing this is possible with datasette if you forgo "canned queries" and just use the raw query endpoint and pass the query sql, along with p1, p2 ... in the request.

I was thinking if there were a way to use some sort of sting function to "unpack" the values and convert them into ints... hm