Module statement 

Statement preparation — building SQL synchronously, so the futures that run it are Send.

Why this module exists

Rainier ORM’s repo:: functions build a sea_query statement and then await the executor with that statement still in scope. sea_query’s statement types hold Rc<dyn Iden>, so the statement is !Send, so the generated future is !Send — and a future that is not Send cannot be awaited inside a handler that a multi-threaded server will tokio::spawn.

The fix is one line upstream (drop(stmt);, or scoping the build, before the await). Until Rainier ORM does that, Rainier prepares statements here instead: every sea_query value is created and dropped inside a synchronous function, which hands back a Prepared holding only String and Value — both Send. The await then touches nothing that is not Send.

What is not re-implemented

The metadata, the dialect rendering, the column types, the row decoding and the shard-routing rule all still come from Rainier ORM. This module is a thin re-statement of repo::’s SQL shapes against the same Entity metadata — not a second ORM. In particular, routing behaviour is deliberately identical to Rainier ORM’s, so a sharded deployment routes the same way whether a query goes through here or through repo:: directly.

Structs

Prepared

A rendered statement: SQL, its ordered bind values, and where to run it.

Functions

count_grouped

SELECT column, COUNT(*) AS cnt … GROUP BY column.

count_matching

SELECT COUNT(*) AS cnt under a Criteria’s filters and joins only.

delete_by_pk

DELETE FROM table WHERE pk = ?.

delete_matching

DELETE FROM table WHERE <criteria>.

insert

INSERT INTO table (…) VALUES (…).

route_for

The route for constraining column to value on E.

routing_key

Reduce a bound value to a u64 routing key.

select_all

SELECT … FROM table — every row.

select_by_column

SELECT … WHERE column = ?, optionally limited.

select_by_pk

SELECT … WHERE pk = ? LIMIT 1.

select_matching

SELECT … under a Criteria’s filters, joins, ordering and paging.

select_pivot

SELECT parent, related FROM pivot WHERE parent IN (…).

shard_key_for_insert

Whether E needs a connector-minted primary key for this row, and the shard key to mint it from.

update

UPDATE table SET … WHERE pk = ? — every non-key column.

update_matching

UPDATE table SET … WHERE <criteria> — only the named columns.

upsert

INSERT … ON CONFLICT (…) DO UPDATE.