graphql-engine/server/src-lib/Hasura/Backends/MSSQL/Execute/Delete.hs

Ignoring revisions in .git-blame-ignore-revs. Click here to bypass and see the normal blame view.

102 lines
4.3 KiB
Haskell
Raw Normal View History

{-# OPTIONS_HADDOCK ignore-exports #-}
-- | Responsible for translating and building an MSSQL execution plan for
-- delete mutations.
--
-- This module is used by "Hasura.Backends.MSSQL.Instances.Execute".
module Hasura.Backends.MSSQL.Execute.Delete
( executeDelete,
)
where
import Database.MSSQL.Transaction qualified as Tx
import Hasura.Backends.MSSQL.Connection
import Hasura.Backends.MSSQL.Execute.QueryTags
import Hasura.Backends.MSSQL.FromIr as TSQL
import Hasura.Backends.MSSQL.FromIr.Constants (tempTableNameDeleted)
import Hasura.Backends.MSSQL.FromIr.Delete qualified as TSQL
import Hasura.Backends.MSSQL.FromIr.MutationResponse
import Hasura.Backends.MSSQL.FromIr.SelectIntoTempTable qualified as TSQL
import Hasura.Backends.MSSQL.Plan
import Hasura.Backends.MSSQL.SQL.Error
import Hasura.Backends.MSSQL.ToQuery as TQ
import Hasura.Backends.MSSQL.Types.Internal as TSQL
import Hasura.Base.Error
import Hasura.EncJSON
Allow backend execution to happen on the base app monad. ### Description Each Backend executes queries against the database in a slightly different stack: Postgres uses its own `TXeT`, MSSQL uses a variant of it, BigQuery is simply in `ExceptT QErr IO`... To accommodate those variations, we had originally introduced an `ExecutionMonad b` type family in `BackendExecute`, allowing each backend to describe its own stack. It was then up to that backend's `BackendTransport` instance to implement running said stack, and converting the result back into our main app monad. However, this was not without complications: `TraceT` is one of them: as it usually needs to be on the top of the stack, converting from one stack to the other implies the use `interpTraceT`, which is quite monstrous. Furthermore, as part of the Entitlement Services work, we're trying to move to a "Services" architecture in which the entire engine runs in one base monad, that delegates features and dependencies to monad constraints; and as a result we'd like to minimize the number of different monad stacks we have to maintain and translate from and to in the codebase. To improve things, this PR changes `ExecutionMonad b` from an _absolute_ stack to a _relative_ one: i.e.: what needs to be stacked on top of our base monad for the execution. In `Transport`, we then only need to pop the top of the stack, and voila. This greatly simplifies the implementation of the backends, as there's no longer any need to do any stack transformation: MySQL's implementation becomes a `runIdentityT`! This also removes most mentions of `TraceT` from the execution code since it's no longer required: we can rely on the base monad's existing `MonadTrace` constraint. To continue encapsulating monadic actions in `DBStepInfo` and avoid threading a bunch of `forall` all over the place, this PR introduces a small local helper: `OnBaseMonad`. One only downside of all this is that this requires adding `MonadBaseControl IO m` constraint all over the place: previously, we would run directly on `IO` and lift, and would therefore not need to bring that constraint all the way. PR-URL: https://github.com/hasura/graphql-engine-mono/pull/7789 GitOrigin-RevId: e9b2e431c5c47fa9851abf87545c0415ff6d1a12
2023-02-09 17:38:33 +03:00
import Hasura.GraphQL.Execute.Backend
import Hasura.Prelude
import Hasura.QueryTags (QueryTagsComment)
import Hasura.RQL.IR
import Hasura.RQL.Types.Backend
import Hasura.RQL.Types.BackendType
import Hasura.RQL.Types.Schema.Options qualified as Options
import Hasura.Session
-- | Executes a Delete IR AST and return results as JSON.
executeDelete ::
(MonadError QErr m, MonadReader QueryTagsComment m) =>
UserInfo ->
Options.StringifyNumbers ->
SourceConfig 'MSSQL ->
AnnDelG 'MSSQL Void (UnpreparedValue 'MSSQL) ->
Allow backend execution to happen on the base app monad. ### Description Each Backend executes queries against the database in a slightly different stack: Postgres uses its own `TXeT`, MSSQL uses a variant of it, BigQuery is simply in `ExceptT QErr IO`... To accommodate those variations, we had originally introduced an `ExecutionMonad b` type family in `BackendExecute`, allowing each backend to describe its own stack. It was then up to that backend's `BackendTransport` instance to implement running said stack, and converting the result back into our main app monad. However, this was not without complications: `TraceT` is one of them: as it usually needs to be on the top of the stack, converting from one stack to the other implies the use `interpTraceT`, which is quite monstrous. Furthermore, as part of the Entitlement Services work, we're trying to move to a "Services" architecture in which the entire engine runs in one base monad, that delegates features and dependencies to monad constraints; and as a result we'd like to minimize the number of different monad stacks we have to maintain and translate from and to in the codebase. To improve things, this PR changes `ExecutionMonad b` from an _absolute_ stack to a _relative_ one: i.e.: what needs to be stacked on top of our base monad for the execution. In `Transport`, we then only need to pop the top of the stack, and voila. This greatly simplifies the implementation of the backends, as there's no longer any need to do any stack transformation: MySQL's implementation becomes a `runIdentityT`! This also removes most mentions of `TraceT` from the execution code since it's no longer required: we can rely on the base monad's existing `MonadTrace` constraint. To continue encapsulating monadic actions in `DBStepInfo` and avoid threading a bunch of `forall` all over the place, this PR introduces a small local helper: `OnBaseMonad`. One only downside of all this is that this requires adding `MonadBaseControl IO m` constraint all over the place: previously, we would run directly on `IO` and lift, and would therefore not need to bring that constraint all the way. PR-URL: https://github.com/hasura/graphql-engine-mono/pull/7789 GitOrigin-RevId: e9b2e431c5c47fa9851abf87545c0415ff6d1a12
2023-02-09 17:38:33 +03:00
m (OnBaseMonad (ExceptT QErr) EncJSON)
executeDelete userInfo stringifyNum sourceConfig deleteOperation = do
queryTags <- ask
preparedDelete <- traverse (prepareValueQuery $ _uiSession userInfo) deleteOperation
Allow backend execution to happen on the base app monad. ### Description Each Backend executes queries against the database in a slightly different stack: Postgres uses its own `TXeT`, MSSQL uses a variant of it, BigQuery is simply in `ExceptT QErr IO`... To accommodate those variations, we had originally introduced an `ExecutionMonad b` type family in `BackendExecute`, allowing each backend to describe its own stack. It was then up to that backend's `BackendTransport` instance to implement running said stack, and converting the result back into our main app monad. However, this was not without complications: `TraceT` is one of them: as it usually needs to be on the top of the stack, converting from one stack to the other implies the use `interpTraceT`, which is quite monstrous. Furthermore, as part of the Entitlement Services work, we're trying to move to a "Services" architecture in which the entire engine runs in one base monad, that delegates features and dependencies to monad constraints; and as a result we'd like to minimize the number of different monad stacks we have to maintain and translate from and to in the codebase. To improve things, this PR changes `ExecutionMonad b` from an _absolute_ stack to a _relative_ one: i.e.: what needs to be stacked on top of our base monad for the execution. In `Transport`, we then only need to pop the top of the stack, and voila. This greatly simplifies the implementation of the backends, as there's no longer any need to do any stack transformation: MySQL's implementation becomes a `runIdentityT`! This also removes most mentions of `TraceT` from the execution code since it's no longer required: we can rely on the base monad's existing `MonadTrace` constraint. To continue encapsulating monadic actions in `DBStepInfo` and avoid threading a bunch of `forall` all over the place, this PR introduces a small local helper: `OnBaseMonad`. One only downside of all this is that this requires adding `MonadBaseControl IO m` constraint all over the place: previously, we would run directly on `IO` and lift, and would therefore not need to bring that constraint all the way. PR-URL: https://github.com/hasura/graphql-engine-mono/pull/7789 GitOrigin-RevId: e9b2e431c5c47fa9851abf87545c0415ff6d1a12
2023-02-09 17:38:33 +03:00
pure $ OnBaseMonad $ mssqlRunReadWrite (_mscExecCtx sourceConfig) (buildDeleteTx preparedDelete stringifyNum queryTags)
-- | Converts a Delete IR AST to a transaction of three delete sql statements.
--
-- A GraphQL delete mutation does two things:
--
-- 1. Deletes rows in a table according to some predicate
-- 2. (Potentially) returns the deleted rows (including relationships) as JSON
--
-- In order to complete these 2 things we need 3 SQL statements:
--
-- 1. @SELECT INTO <temp_table> WHERE <false>@ - creates a temporary table
-- with the same schema as the original table in which we'll store the deleted rows
-- from the table we are deleting
-- 2. @DELETE FROM with OUTPUT@ - deletes the rows from the table and inserts the
-- deleted rows to the temporary table from (1)
-- 3. @SELECT@ - constructs the @returning@ query from the temporary table, including
-- relationships with other tables.
buildDeleteTx ::
Allow backend execution to happen on the base app monad. ### Description Each Backend executes queries against the database in a slightly different stack: Postgres uses its own `TXeT`, MSSQL uses a variant of it, BigQuery is simply in `ExceptT QErr IO`... To accommodate those variations, we had originally introduced an `ExecutionMonad b` type family in `BackendExecute`, allowing each backend to describe its own stack. It was then up to that backend's `BackendTransport` instance to implement running said stack, and converting the result back into our main app monad. However, this was not without complications: `TraceT` is one of them: as it usually needs to be on the top of the stack, converting from one stack to the other implies the use `interpTraceT`, which is quite monstrous. Furthermore, as part of the Entitlement Services work, we're trying to move to a "Services" architecture in which the entire engine runs in one base monad, that delegates features and dependencies to monad constraints; and as a result we'd like to minimize the number of different monad stacks we have to maintain and translate from and to in the codebase. To improve things, this PR changes `ExecutionMonad b` from an _absolute_ stack to a _relative_ one: i.e.: what needs to be stacked on top of our base monad for the execution. In `Transport`, we then only need to pop the top of the stack, and voila. This greatly simplifies the implementation of the backends, as there's no longer any need to do any stack transformation: MySQL's implementation becomes a `runIdentityT`! This also removes most mentions of `TraceT` from the execution code since it's no longer required: we can rely on the base monad's existing `MonadTrace` constraint. To continue encapsulating monadic actions in `DBStepInfo` and avoid threading a bunch of `forall` all over the place, this PR introduces a small local helper: `OnBaseMonad`. One only downside of all this is that this requires adding `MonadBaseControl IO m` constraint all over the place: previously, we would run directly on `IO` and lift, and would therefore not need to bring that constraint all the way. PR-URL: https://github.com/hasura/graphql-engine-mono/pull/7789 GitOrigin-RevId: e9b2e431c5c47fa9851abf87545c0415ff6d1a12
2023-02-09 17:38:33 +03:00
(MonadIO m) =>
AnnDel 'MSSQL ->
Options.StringifyNumbers ->
QueryTagsComment ->
Allow backend execution to happen on the base app monad. ### Description Each Backend executes queries against the database in a slightly different stack: Postgres uses its own `TXeT`, MSSQL uses a variant of it, BigQuery is simply in `ExceptT QErr IO`... To accommodate those variations, we had originally introduced an `ExecutionMonad b` type family in `BackendExecute`, allowing each backend to describe its own stack. It was then up to that backend's `BackendTransport` instance to implement running said stack, and converting the result back into our main app monad. However, this was not without complications: `TraceT` is one of them: as it usually needs to be on the top of the stack, converting from one stack to the other implies the use `interpTraceT`, which is quite monstrous. Furthermore, as part of the Entitlement Services work, we're trying to move to a "Services" architecture in which the entire engine runs in one base monad, that delegates features and dependencies to monad constraints; and as a result we'd like to minimize the number of different monad stacks we have to maintain and translate from and to in the codebase. To improve things, this PR changes `ExecutionMonad b` from an _absolute_ stack to a _relative_ one: i.e.: what needs to be stacked on top of our base monad for the execution. In `Transport`, we then only need to pop the top of the stack, and voila. This greatly simplifies the implementation of the backends, as there's no longer any need to do any stack transformation: MySQL's implementation becomes a `runIdentityT`! This also removes most mentions of `TraceT` from the execution code since it's no longer required: we can rely on the base monad's existing `MonadTrace` constraint. To continue encapsulating monadic actions in `DBStepInfo` and avoid threading a bunch of `forall` all over the place, this PR introduces a small local helper: `OnBaseMonad`. One only downside of all this is that this requires adding `MonadBaseControl IO m` constraint all over the place: previously, we would run directly on `IO` and lift, and would therefore not need to bring that constraint all the way. PR-URL: https://github.com/hasura/graphql-engine-mono/pull/7789 GitOrigin-RevId: e9b2e431c5c47fa9851abf87545c0415ff6d1a12
2023-02-09 17:38:33 +03:00
Tx.TxET QErr m EncJSON
buildDeleteTx deleteOperation stringifyNum queryTags = do
let withAlias = "with_alias"
createInsertedTempTableQuery =
toQueryFlat
$ TQ.fromSelectIntoTempTable
$ TSQL.toSelectIntoTempTable tempTableNameDeleted (_adTable deleteOperation) (_adAllCols deleteOperation) RemoveConstraints
-- Create a temp table
Tx.unitQueryE defaultMSSQLTxErrorHandler (createInsertedTempTableQuery `withQueryTags` queryTags)
let deleteQuery = TQ.fromDelete <$> TSQL.fromDelete deleteOperation
deleteQueryValidated <- toQueryFlat . qwdQuery <$> runFromIrErrorOnCTEs deleteQuery
-- Execute DELETE statement
Tx.unitQueryE mutationMSSQLTxErrorHandler (deleteQueryValidated `withQueryTags` queryTags)
mutationOutputSelect <- qwdQuery <$> runFromIrUseCTEs (mkMutationOutputSelect stringifyNum withAlias $ _adOutput deleteOperation)
let withSelect =
emptySelect
{ selectProjections = [StarProjection],
selectFrom = Just $ FromTempTable $ Aliased tempTableNameDeleted "deleted_alias"
}
finalMutationOutputSelect = mutationOutputSelect {selectWith = Just $ With $ pure $ Aliased (CTESelect withSelect) withAlias}
mutationOutputSelectQuery = toQueryFlat $ TQ.fromSelect finalMutationOutputSelect
-- Execute SELECT query and fetch mutation response
result <- encJFromText <$> Tx.singleRowQueryE defaultMSSQLTxErrorHandler (mutationOutputSelectQuery `withQueryTags` queryTags)
-- delete the temporary table
let dropDeletedTempTableQuery = toQueryFlat $ dropTempTableQuery tempTableNameDeleted
Tx.unitQueryE defaultMSSQLTxErrorHandler (dropDeletedTempTableQuery `withQueryTags` queryTags)
-- return results
pure result