graphql-engine/server/src-lib/Hasura/Backends/Postgres/Translate/Select/Aggregate.hs

-- | This module defines the top-level translation functions pertaining to
-- queries that use aggregation (i.e., /not/ so-called "simple" selects) into
-- Postgres AST.
module Hasura.Backends.Postgres.Translate.Select.Aggregate
  ( mkAggregateSelect,
    selectAggregateQuerySQL,
  )
where

import Control.Monad.Writer.Strict (runWriter)
import Database.PG.Query (Query, fromBuilder)
import Hasura.Backends.Postgres.SQL.DML (BoolExp (BELit), Select)
import Hasura.Backends.Postgres.SQL.RenameIdentifiers (renameIdentifiers)
import Hasura.Backends.Postgres.SQL.Types (IsIdentifier (toIdentifier))
import Hasura.Backends.Postgres.Translate.Select.AnnotatedFieldJSON
import Hasura.Backends.Postgres.Translate.Select.Internal.GenerateSelect (generateSQLSelectFromArrayNode)
import Hasura.Backends.Postgres.Translate.Select.Internal.Process (processAnnAggregateSelect)
import Hasura.Backends.Postgres.Translate.Types (MultiRowSelectNode (MultiRowSelectNode), SelectNode (SelectNode), SourcePrefixes (SourcePrefixes))
import Hasura.Prelude
import Hasura.RQL.IR.Select (AnnAggregateSelect, AnnSelectG (_asnStrfyNum))
import Hasura.RQL.Types.Backend (Backend)
import Hasura.RQL.Types.Common (FieldName (FieldName))
import Hasura.SQL.Backend (BackendType (Postgres))
import Hasura.SQL.Types (ToSQL (toSQL))

-- | Translates IR to Postgres queries for aggregated SELECTs.
--
-- See 'mkAggregateSelect' for the Postgres AST.
selectAggregateQuerySQL ::
  forall pgKind.
  (Backend ('Postgres pgKind), PostgresAnnotatedFieldJSON pgKind) =>
  AnnAggregateSelect ('Postgres pgKind) ->
  Query
selectAggregateQuerySQL = fromBuilder . toSQL . mkAggregateSelect

-- | We process aggregate queries differently because the types of aggregate queries are different.
--   In the '_asnFields' field of an 'AnnSelectG', we will have a 'TableAggregateFieldG' instead
--   of an 'AnnFieldG'.
mkAggregateSelect ::
  forall pgKind.
  (Backend ('Postgres pgKind), PostgresAnnotatedFieldJSON pgKind) =>
  AnnAggregateSelect ('Postgres pgKind) ->
  Select
mkAggregateSelect annAggSel =
  let ((selectSource, nodeExtractors, topExtractor), joinTree) =
        runWriter $
          flip runReaderT strfyNum $
            processAnnAggregateSelect sourcePrefixes rootFieldName annAggSel
      -- select the relevant columns and subquery we want to aggregate
      selectNode = SelectNode nodeExtractors joinTree
      -- aggregate the results into a top-level return value
      arrayNode = MultiRowSelectNode [topExtractor] selectNode
   in renameIdentifiers $
        generateSQLSelectFromArrayNode selectSource arrayNode $
          BELit True
  where
    strfyNum = _asnStrfyNum annAggSel
    rootFieldName = FieldName "root"
    rootIdentifier = toIdentifier rootFieldName
    sourcePrefixes = SourcePrefixes rootIdentifier rootIdentifier