Notice that sbroad function are placed into global [_G](https://www.lua.org/pil/14.html), so you can call sbroad functions at any place of app. Let's have a look at sbroad functions:
execute the parametrized sql query. Currently query supports only _SELECT_, _INSERT_ and _EXPLAIN_ statements. For details see [SQL feature taxonomy](https://git.picodata.io/picodata/picodata/sbroad/-/blob/main/doc/sql/feature_taxonomy.md).
The `query` function executes the sql query from the `sql` parameter.
**Parameters**:
* sql_query (string) -- sql query to execute on cluster
* params (array) -- params for sql query. The number and order of the arguments correspond to the `?` signs inside `sql_query`.
The `insert_record` function detects the bucket from `values_map` parameter and calls `insert_map` function on the bucket storage.
**Return**:
* result of query (table|nil)
* error
The `storage` role has the `insert_map(space, values_map)` function, which inserts map of values `values_map` into `space`.
You can find more examples inside [integration tests](https://git.picodata.io/picodata/picodata/sbroad/-/tree/main/sbroad-cartridge/test_app/test/integration). To run it use
* value (string|tuple|table) -- value on the basis of which to calculate
* space_name (string) -- name of space.
**Restrictions**:
* String type value does not need space_name
* Tuple type value (array or box.tuple) needs `space_name`. The number and order of the elements must be corresponding to space format or less by one (skip bucket_id field).
* Table type value needs `space_name`. Table must contain all space sharding_key elements.
**Return**:
* result (number|nil) -- bucket_id
* error
**Example**:
```
sbroad.calculate_bucket_id("1123")
---
- 360
...
sbroad.calculate_bucket_id({ id = 1, name = "123" }, "testing_space"})
execute the parametrized sql query in the same way as [sbroad.execute](#sbroad-execute), but also start trace of query.
**Parameters**:
* sql_query (string) -- sql query to execute on cluster
* params (array) -- params for sql query. The number and order of the arguments correspond to the `?` signs inside `sql_query`
* carrier (table|box.NULL) -- tracer params that can be represented as [W3C format](https://www.w3.org/TR/trace-context/#traceparent-header) or as [Jaegger format](https://www.jaegertracing.io/docs/1.38/client-libraries/#propagation-format)
`Sbroad` library uses internal lua functions in the cartridge executor and preloads them with `load_lua_extra_function` call in the `init` cartridge function.
As the `sbroad` library caches the cluster cartridge schema internally, any `sbroad` function that is called checks the internal cluster schema, and if that is empty it loads the schema from the main app. If the app schema was updated then the internal cache needs to be cleared. To clear the cache we need to add the `invalidate_cached_schema` call to the `apply_config` cartridge function.
## Local load testing
## <a name="load-test"></a>Local load testing
More information in [stress-test](../stress-test) folder.
