mirror of
https://github.com/hasura/graphql-engine.git
synced 2024-12-18 04:51:35 +03:00
e8e4f30dd6
Remote relationships are now supported on SQL Server and BigQuery. The major change though is the re-architecture of remote join execution logic. Prior to this PR, each backend is responsible for processing the remote relationships that are part of their AST. This is not ideal as there is nothing specific about a remote join's execution that ties it to a backend. The only backend specific part is whether or not the specification of the remote relationship is valid (i.e, we'll need to validate whether the scalars are compatible). The approach now changes to this: 1. Before delegating the AST to the backend, we traverse the AST, collect all the remote joins while modifying the AST to add necessary join fields where needed. 1. Once the remote joins are collected from the AST, the database call is made to fetch the response. The necessary data for the remote join(s) is collected from the database's response and one or more remote schema calls are constructed as necessary. 1. The remote schema calls are then executed and the data from the database and from the remote schemas is joined to produce the final response. ### Known issues 1. Ideally the traversal of the IR to collect remote joins should return an AST which does not include remote join fields. This operation can be type safe but isn't taken up as part of the PR. 1. There is a lot of code duplication between `Transport/HTTP.hs` and `Transport/Websocket.hs` which needs to be fixed ASAP. This too hasn't been taken up by this PR. 1. The type which represents the execution plan is only modified to handle our current remote joins and as such it will have to be changed to accommodate general remote joins. 1. Use of lenses would have reduced the boilerplate code to collect remote joins from the base AST. 1. The current remote join logic assumes that the join columns of a remote relationship appear with their names in the database response. This however is incorrect as they could be aliased. This can be taken up by anyone, I've left a comment in the code. ### Notes to the reviewers I think it is best reviewed commit by commit. 1. The first one is very straight forward. 1. The second one refactors the remote join execution logic but other than moving things around, it doesn't change the user facing functionality. This moves Postgres specific parts to `Backends/Postgres` module from `Execute`. Some IR related code to `Hasura.RQL.IR` module. Simplifies various type class function signatures as a backend doesn't have to handle remote joins anymore 1. The third one fixes partial case matches that for some weird reason weren't shown as warnings before this refactor 1. The fourth one generalizes the validation logic of remote relationships and implements `scalarTypeGraphQLName` function on SQL Server and BigQuery which is used by the validation logic. This enables remote relationships on BigQuery and SQL Server. https://github.com/hasura/graphql-engine-mono/pull/1497 GitOrigin-RevId: 77dd8eed326602b16e9a8496f52f46d22b795598
263 lines
8.7 KiB
Haskell
263 lines
8.7 KiB
Haskell
module Hasura.Backends.Postgres.Execute.Mutation
|
|
( MutationRemoteJoinCtx
|
|
, MutateResp(..)
|
|
--
|
|
, execDeleteQuery
|
|
, execInsertQuery
|
|
, execUpdateQuery
|
|
--
|
|
, executeMutationOutputQuery
|
|
, mutateAndFetchCols
|
|
) where
|
|
|
|
import Hasura.Prelude
|
|
|
|
import qualified Data.Sequence as DS
|
|
import qualified Database.PG.Query as Q
|
|
import qualified Network.HTTP.Client as HTTP
|
|
import qualified Network.HTTP.Types as N
|
|
|
|
import Data.Aeson
|
|
|
|
import qualified Hasura.Backends.Postgres.SQL.DML as S
|
|
|
|
import Hasura.Backends.Postgres.Connection
|
|
import Hasura.Backends.Postgres.SQL.Types
|
|
import Hasura.Backends.Postgres.SQL.Value
|
|
import Hasura.Backends.Postgres.Translate.Delete
|
|
import Hasura.Backends.Postgres.Translate.Insert
|
|
import Hasura.Backends.Postgres.Translate.Mutation
|
|
import Hasura.Backends.Postgres.Translate.Returning
|
|
import Hasura.Backends.Postgres.Translate.Select
|
|
import Hasura.Backends.Postgres.Translate.Update
|
|
import Hasura.Base.Error
|
|
import Hasura.Base.Instances ()
|
|
import Hasura.EncJSON
|
|
import Hasura.RQL.DML.Internal
|
|
import Hasura.RQL.IR.Delete
|
|
import Hasura.RQL.IR.Insert
|
|
import Hasura.RQL.IR.Returning
|
|
import Hasura.RQL.IR.Select
|
|
import Hasura.RQL.IR.Update
|
|
import Hasura.RQL.Types
|
|
import Hasura.SQL.Types
|
|
import Hasura.Session
|
|
|
|
|
|
data MutateResp (b :: BackendType) a
|
|
= MutateResp
|
|
{ _mrAffectedRows :: !Int
|
|
, _mrReturningColumns :: ![ColumnValues b a]
|
|
} deriving (Generic)
|
|
deriving instance (Backend b, Show a) => Show (MutateResp b a)
|
|
deriving instance (Backend b, Eq a) => Eq (MutateResp b a)
|
|
|
|
instance (Backend b, ToJSON a) => ToJSON (MutateResp b a) where
|
|
toJSON = genericToJSON hasuraJSON
|
|
|
|
instance (Backend b, FromJSON a) => FromJSON (MutateResp b a) where
|
|
parseJSON = genericParseJSON hasuraJSON
|
|
|
|
|
|
type MutationRemoteJoinCtx = (HTTP.Manager, [N.Header], UserInfo)
|
|
|
|
data Mutation (b :: BackendType)
|
|
= Mutation
|
|
{ _mTable :: !QualifiedTable
|
|
, _mQuery :: !(MutationCTE, DS.Seq Q.PrepArg)
|
|
, _mOutput :: !(MutationOutput b)
|
|
, _mCols :: ![ColumnInfo b]
|
|
, _mStrfyNum :: !Bool
|
|
}
|
|
|
|
mkMutation
|
|
:: UserInfo
|
|
-> QualifiedTable
|
|
-> (MutationCTE, DS.Seq Q.PrepArg)
|
|
-> MutationOutput ('Postgres pgKind)
|
|
-> [ColumnInfo ('Postgres pgKind)]
|
|
-> Bool
|
|
-> Mutation ('Postgres pgKind)
|
|
mkMutation _userInfo table query output allCols strfyNum =
|
|
Mutation table query output allCols strfyNum
|
|
|
|
runMutation
|
|
::
|
|
( MonadTx m
|
|
, Backend ('Postgres pgKind)
|
|
, PostgresAnnotatedFieldJSON pgKind
|
|
)
|
|
=> Mutation ('Postgres pgKind)
|
|
-> m EncJSON
|
|
runMutation mut =
|
|
bool (mutateAndReturn mut) (mutateAndSel mut) $
|
|
hasNestedFld $ _mOutput mut
|
|
|
|
mutateAndReturn
|
|
::
|
|
( MonadTx m
|
|
, Backend ('Postgres pgKind)
|
|
, PostgresAnnotatedFieldJSON pgKind
|
|
)
|
|
=> Mutation ('Postgres pgKind)
|
|
-> m EncJSON
|
|
mutateAndReturn (Mutation qt (cte, p) mutationOutput allCols strfyNum) =
|
|
executeMutationOutputQuery qt allCols Nothing cte mutationOutput strfyNum (toList p)
|
|
|
|
|
|
execUpdateQuery
|
|
::
|
|
( MonadTx m
|
|
, Backend ('Postgres pgKind)
|
|
, PostgresAnnotatedFieldJSON pgKind
|
|
)
|
|
=> Bool
|
|
-> UserInfo
|
|
-> (AnnUpd ('Postgres pgKind), DS.Seq Q.PrepArg)
|
|
-> m EncJSON
|
|
execUpdateQuery strfyNum userInfo (u, p) =
|
|
runMutation $ mkMutation userInfo (uqp1Table u) (MCCheckConstraint updateCTE, p)
|
|
(uqp1Output u) (uqp1AllCols u) strfyNum
|
|
where
|
|
updateCTE = mkUpdateCTE u
|
|
|
|
execDeleteQuery
|
|
::
|
|
( MonadTx m
|
|
, Backend ('Postgres pgKind)
|
|
, PostgresAnnotatedFieldJSON pgKind
|
|
)
|
|
=> Bool
|
|
-> UserInfo
|
|
-> (AnnDel ('Postgres pgKind), DS.Seq Q.PrepArg)
|
|
-> m EncJSON
|
|
execDeleteQuery strfyNum remoteJoinCtx (u, p) =
|
|
runMutation $ mkMutation remoteJoinCtx (dqp1Table u) (MCDelete delete, p)
|
|
(dqp1Output u) (dqp1AllCols u) strfyNum
|
|
where
|
|
delete = mkDelete u
|
|
|
|
execInsertQuery
|
|
:: ( MonadTx m
|
|
, Backend ('Postgres pgKind)
|
|
, PostgresAnnotatedFieldJSON pgKind
|
|
)
|
|
=> Bool
|
|
-> UserInfo
|
|
-> (InsertQueryP1 ('Postgres pgKind), DS.Seq Q.PrepArg)
|
|
-> m EncJSON
|
|
execInsertQuery strfyNum userInfo (u, p) =
|
|
runMutation
|
|
$ mkMutation userInfo (iqp1Table u) (MCCheckConstraint insertCTE, p)
|
|
(iqp1Output u) (iqp1AllCols u) strfyNum
|
|
where
|
|
insertCTE = mkInsertCTE u
|
|
|
|
|
|
|
|
{- Note: [Prepared statements in Mutations]
|
|
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
|
|
The SQL statements we generate for mutations seem to include the actual values
|
|
in the statements in some cases which pretty much makes them unfit for reuse
|
|
(Handling relationships in the returning clause is the source of this
|
|
complexity). Further, `PGConn` has an internal cache which maps a statement to
|
|
a 'prepared statement id' on Postgres. As we prepare more and more single-use
|
|
SQL statements we end up leaking memory both on graphql-engine and Postgres
|
|
till the connection is closed. So a simpler but very crude fix is to not use
|
|
prepared statements for mutations. The performance of insert mutations
|
|
shouldn't be affected but updates and delete mutations with complex boolean
|
|
conditions **might** see some degradation.
|
|
-}
|
|
|
|
mutateAndSel
|
|
:: forall pgKind m
|
|
. ( MonadTx m
|
|
, Backend ('Postgres pgKind)
|
|
, PostgresAnnotatedFieldJSON pgKind
|
|
)
|
|
=> Mutation ('Postgres pgKind)
|
|
-> m EncJSON
|
|
mutateAndSel (Mutation qt q mutationOutput allCols strfyNum) = do
|
|
-- Perform mutation and fetch unique columns
|
|
MutateResp _ columnVals <- liftTx $ mutateAndFetchCols qt allCols q strfyNum
|
|
select <- mkSelectExpFromColumnValues qt allCols columnVals
|
|
-- Perform select query and fetch returning fields
|
|
executeMutationOutputQuery qt allCols Nothing
|
|
(MCSelectValues select) mutationOutput strfyNum []
|
|
|
|
withCheckPermission :: (MonadError QErr m) => m (a, Bool) -> m a
|
|
withCheckPermission sqlTx = do
|
|
(rawResponse, checkConstraint) <- sqlTx
|
|
unless checkConstraint $ throw400 PermissionError $
|
|
"check constraint of an insert/update permission has failed"
|
|
pure rawResponse
|
|
|
|
executeMutationOutputQuery
|
|
:: forall pgKind m
|
|
. ( MonadTx m
|
|
, Backend ('Postgres pgKind)
|
|
, PostgresAnnotatedFieldJSON pgKind
|
|
)
|
|
=> QualifiedTable
|
|
-> [ColumnInfo ('Postgres pgKind)]
|
|
-> Maybe Int
|
|
-> MutationCTE
|
|
-> MutationOutput ('Postgres pgKind)
|
|
-> Bool
|
|
-> [Q.PrepArg] -- ^ Prepared params
|
|
-> m EncJSON
|
|
executeMutationOutputQuery qt allCols preCalAffRows cte mutOutput strfyNum prepArgs = do
|
|
let queryTx :: Q.FromRes a => m a
|
|
queryTx = do
|
|
let selectWith = mkMutationOutputExp qt allCols preCalAffRows cte mutOutput strfyNum
|
|
query = Q.fromBuilder $ toSQL selectWith
|
|
-- See Note [Prepared statements in Mutations]
|
|
liftTx (Q.rawQE dmlTxErrorHandler query prepArgs False)
|
|
|
|
if checkPermissionRequired cte
|
|
then withCheckPermission $ Q.getRow <$> queryTx
|
|
else (runIdentity . Q.getRow) <$> queryTx
|
|
|
|
mutateAndFetchCols
|
|
:: forall pgKind
|
|
. (Backend ('Postgres pgKind), PostgresAnnotatedFieldJSON pgKind)
|
|
=> QualifiedTable
|
|
-> [ColumnInfo ('Postgres pgKind)]
|
|
-> (MutationCTE, DS.Seq Q.PrepArg)
|
|
-> Bool
|
|
-> Q.TxE QErr (MutateResp ('Postgres pgKind) TxtEncodedVal)
|
|
mutateAndFetchCols qt cols (cte, p) strfyNum = do
|
|
let mutationTx :: Q.FromRes a => Q.TxE QErr a
|
|
mutationTx =
|
|
-- See Note [Prepared statements in Mutations]
|
|
Q.rawQE dmlTxErrorHandler sqlText (toList p) False
|
|
|
|
if checkPermissionRequired cte
|
|
then withCheckPermission $ (first Q.getAltJ . Q.getRow) <$> mutationTx
|
|
else (Q.getAltJ . runIdentity . Q.getRow) <$> mutationTx
|
|
where
|
|
aliasIdentifier = Identifier $ qualifiedObjectToText qt <> "__mutation_result"
|
|
tabFrom = FromIdentifier aliasIdentifier
|
|
tabPerm = TablePerm annBoolExpTrue Nothing
|
|
selFlds = flip map cols $
|
|
\ci -> (fromCol @('Postgres pgKind) $ pgiColumn ci, mkAnnColumnFieldAsText ci)
|
|
|
|
sqlText = Q.fromBuilder $ toSQL selectWith
|
|
selectWith = S.SelectWith [(S.Alias aliasIdentifier, getMutationCTE cte)] select
|
|
select = S.mkSelect { S.selExtr = S.Extractor extrExp Nothing
|
|
: bool [] [S.Extractor checkErrExp Nothing] (checkPermissionRequired cte)
|
|
}
|
|
checkErrExp = mkCheckErrorExp aliasIdentifier
|
|
extrExp = S.applyJsonBuildObj
|
|
[ S.SELit "affected_rows", affRowsSel
|
|
, S.SELit "returning_columns", colSel
|
|
]
|
|
|
|
affRowsSel = S.SESelect $
|
|
S.mkSelect
|
|
{ S.selExtr = [S.Extractor S.countStar Nothing]
|
|
, S.selFrom = Just $ S.FromExp [S.FIIdentifier aliasIdentifier]
|
|
}
|
|
colSel = S.SESelect $ mkSQLSelect JASMultipleRows $
|
|
AnnSelectG selFlds tabFrom tabPerm noSelectArgs strfyNum
|