mirror of
https://github.com/hasura/graphql-engine.git
synced 2024-12-15 01:12:56 +03:00
Group mutation nodes by connector (#406)
## Description In order to implement transactions for the connectors that allow it, we want to group mutation commands by connector. Then, we can execute a group in a single transaction if possible, else just loop through it as we always have. Grouping by connector also means we can access things like capabilities for whole batches at a time. ## Changelog - Add a changelog entry (in the "Changelog entry" section below) if the changes in this PR have any user-facing impact. See [changelog guide](https://github.com/hasura/graphql-engine-mono/wiki/Changelog-Guide). - If no changelog is required ignore/remove this section and add a `no-changelog-required` label to the PR. ### Product _(Select all products this will be available in)_ - [ ] community-edition - [ ] cloud <!-- product : end : DO NOT REMOVE --> ### Type <!-- See changelog structure: https://github.com/hasura/graphql-engine-mono/wiki/Changelog-Guide#structure-of-our-changelog --> _(Select only one. In case of multiple, choose the most appropriate)_ - [ ] highlight - [ ] enhancement - [ ] bugfix - [ ] behaviour-change - [ ] performance-enhancement - [ ] security-fix <!-- type : end : DO NOT REMOVE --> ### Changelog entry <!-- - Add a user understandable changelog entry - Include all details needed to understand the change. Try including links to docs or issues if relevant - For Highlights start with a H4 heading (#### <entry title>) - Get the changelog entry reviewed by your team --> _Replace with changelog entry_ <!-- changelog-entry : end : DO NOT REMOVE --> <!-- changelog : end : DO NOT REMOVE --> V3_GIT_ORIGIN_REV_ID: 928715b78b214a3ac2452e8d1e5f33c87b20de33
This commit is contained in:
parent
2ea7dbde76
commit
fe6e99762e
@ -153,7 +153,8 @@ pub(crate) async fn explain_mutation_plan(
|
||||
));
|
||||
}
|
||||
|
||||
for (alias, ndc_mutation_execution) in mutation_plan.nodes {
|
||||
for (_, mutation_group) in mutation_plan.nodes {
|
||||
for (alias, ndc_mutation_execution) in mutation_group {
|
||||
let sequence_steps = get_execution_steps(
|
||||
http_client,
|
||||
alias,
|
||||
@ -165,6 +166,7 @@ pub(crate) async fn explain_mutation_plan(
|
||||
.await;
|
||||
root_steps.push(Box::new(types::Step::Sequence(sequence_steps)));
|
||||
}
|
||||
}
|
||||
|
||||
// simplify the steps
|
||||
match NonEmpty::from_vec(root_steps) {
|
||||
|
@ -27,8 +27,17 @@ use crate::schema::GDS;
|
||||
|
||||
pub type QueryPlan<'n, 's, 'ir> = IndexMap<ast::Alias, NodeQueryPlan<'n, 's, 'ir>>;
|
||||
|
||||
/// Unlike a query, the root nodes of a mutation aren't necessarily independent. Specifically, the
|
||||
/// GraphQL specification says that each root mutation must be executed sequentially. Moreover, if
|
||||
/// we want to, say, insert a parent _and_ children in one query, we want the ability to make
|
||||
/// transactional requests. In a mutation plan, we group nodes by connector, allowing us to issue
|
||||
/// transactional commands to connectors whose capabilities allow for transactional mutations.
|
||||
/// Otherwise, we can just send them one-by-one (though still sequentially).
|
||||
pub struct MutationPlan<'n, 's, 'ir> {
|
||||
pub nodes: IndexMap<ast::Alias, NDCMutationExecution<'n, 's, 'ir>>,
|
||||
pub nodes: IndexMap<
|
||||
resolved::data_connector::DataConnectorLink,
|
||||
IndexMap<ast::Alias, NDCMutationExecution<'n, 's, 'ir>>,
|
||||
>,
|
||||
pub type_names: IndexMap<ast::Alias, ast::TypeName>,
|
||||
}
|
||||
|
||||
@ -184,9 +193,13 @@ pub fn generate_request_plan<'n, 's, 'ir>(
|
||||
.insert(alias.clone(), type_name.clone());
|
||||
}
|
||||
root_field::MutationRootField::ProcedureBasedCommand { selection_set, ir } => {
|
||||
let plan = plan_mutation(selection_set, ir)?;
|
||||
|
||||
mutation_plan
|
||||
.nodes
|
||||
.insert(alias.clone(), plan_mutation(selection_set, ir)?);
|
||||
.entry(plan.data_connector.clone())
|
||||
.or_default()
|
||||
.insert(alias.clone(), plan);
|
||||
}
|
||||
};
|
||||
|
||||
@ -728,12 +741,14 @@ pub async fn execute_mutation_plan<'n, 's, 'ir>(
|
||||
));
|
||||
}
|
||||
|
||||
for (alias, field_plan) in mutation_plan.nodes.into_iter() {
|
||||
for (_, mutation_group) in mutation_plan.nodes {
|
||||
for (alias, field_plan) in mutation_group {
|
||||
executed_root_fields.push((
|
||||
alias,
|
||||
execute_mutation_field_plan(http_client, field_plan, project_id.clone()).await,
|
||||
));
|
||||
}
|
||||
}
|
||||
|
||||
for executed_root_field in executed_root_fields.into_iter() {
|
||||
let (alias, root_field) = executed_root_field;
|
||||
|
@ -26,6 +26,12 @@ pub struct DataConnectorLink {
|
||||
pub headers: SerializableHeaderMap,
|
||||
}
|
||||
|
||||
impl std::hash::Hash for DataConnectorLink {
|
||||
fn hash<H>(&self, h: &mut H) where H: std::hash::Hasher {
|
||||
self.name.hash(h)
|
||||
}
|
||||
}
|
||||
|
||||
impl DataConnectorLink {
|
||||
pub(crate) fn new(
|
||||
name: Qualified<DataConnectorName>,
|
||||
|
Loading…
Reference in New Issue
Block a user