1.5.0.0 — 2024-03-19

Removed

• Removed nullaryFunction. Instead function can be called with (). (#258)

Added

• Support PostgreSQL's inet type (which maps to the Haskell NetAddr IP type). (#227)

• Rel8.materialize and Rel8.Tabulate.materialize, which add a materialization/optimisation fence to SELECT statements by binding a query to a WITH subquery. Note that explicitly materialized common table expressions are only supported in PostgreSQL 12 an higher. (#180) (#284)

• Rel8.head, Rel8.headExpr, Rel8.last, Rel8.lastExpr for accessing the first/last elements of ListTables and arrays. We have also added variants for NonEmptyTables/non-empty arrays with the 1 suffix (e.g., head1). (#245)

• Rel8 now has extensive support for WITH statements and data-modifying statements (https://www.postgresql.org/docs/current/queries-with.html#QUERIES-WITH-MODIFYING).

  This work offers a lot of new power to Rel8. One new possibility is "moving" rows between tables, for example to archive rows in one table into a log table:

  import Rel8
  
  archive :: Statement ()
  archive = do
    deleted <-
      delete Delete
        { from = mainTable
        , using = pure ()
        , deleteWhere = \foo -> fooId foo ==. lit 123
        , returning = Returning id
        }
  
    insert Insert
      { into = archiveTable
      , rows = deleted
      , onConflict = DoNothing
      , returning = NoReturninvg
      }
  

  This Statement will compile to a single SQL statement - essentially:

  WITH deleted_rows (DELETE FROM main_table WHERE id = 123 RETURNING *)
  INSERT INTO archive_table SELECT * FROM deleted_rows
  

  This feature is a significant performant improvement, as it avoids an entire roundtrip.

  This change has necessitated a change to how a SELECT statement is ran: select now will now produce a Rel8.Statement, which you have to run to turn it into a Hasql Statement. Rel8 offers a variety of run functions depending on how many rows need to be returned - see the various family of run functions in Rel8's documentation for more.

  #250

• Rel8.loop and Rel8.loopDistinct, which allow writing WITH .. RECURSIVE queries. (#180)

• Added the QualifiedName type for named PostgreSQL objects (tables, views, functions, operators, sequences, etc.) that can optionally be qualified by a schema, including an IsString instance. (#257) (#263)

• Added queryFunction for SELECTing from table-returning functions such as jsonb_to_recordset. (#241)

• TypeName record, which gives a richer representation of the components of a PostgreSQL type name (name, schema, modifiers, scalar/array). (#263)

• Rel8.length and Rel8.lengthExpr for getting the length ListTables and arrays. We have also added variants for NonEmptyTables/non-empty arrays with the 1 suffix (e.g., length1). (#268)

• Added aggregators listCat and nonEmptyCat for folding a collection of lists into a single list by concatenation. (#270)

• DBType instance for Fixed that would map (e.g.) Micro to numeric(1000, 6) and Pico to numeric(1000, 12). (#280)

• aggregationFunction, which allows custom aggregation functions to be used. (#283)

• Add support for ordered-set aggregation functions, including mode, percentile, percentileContinuous, hypotheticalRank, hypotheticalDenseRank, hypotheticalPercentRank and hypotheticalCumeDist. (#282)

• Added index, index1, indexExpr, and index1Expr functions for extracting individual elements from ListTables and NonEmptyTables. (#285)

• Rel8 now supports GHC 9.8. (#299)

Changed

• Rel8's API regarding aggregation has changed significantly, and is now a closer match to Opaleye.

  The previous aggregation API had aggregate transform a Table from the Aggregate context back into the Expr context:

  myQuery = aggregate do
    a <- each tableA
    return $ liftF2 (,) (sum (foo a)) (countDistinct (bar a))
  

  This API seemed convenient, but has some significant shortcomings. The new API requires an explicit Aggregator be passed to aggregate:

  myQuery = aggregate (liftA2 (,) (sumOn foo) (countDistinctOn bar)) do
    each tableA
  

  For more details, see #235

• TypeInformation's decoder field has changed. Instead of taking a Hasql.Decoder, it now takes a Rel8.Decoder, which itself is comprised of a Hasql.Decoder and an attoparsec Parser. This is necessitated by the fix for #168; we generally decode things in PostgreSQL's binary format (using a Hasql.Decoder), but for nested arrays we now get things in PostgreSQL's text format (for which we need an attoparsec Parser), so must have both. Most DBType instances that use mapTypeInformation or ParseTypeInformation, or DerivingVia helpers like ReadShow, JSONBEncoded, Enum and Composite are unaffected by this change. (#243)

• The schema field from TableSchema has been removed and the name field changed from String to QualifiedName. (#257)

• nextval, function and binaryOperator now take a QualifiedName instead of a String. (#262)

• function has been changed to accept a single argument (as opposed to variadic arguments). (#258)

• TypeInformation's typeName parameter from String to TypeName. (#263)

• DBEnum's enumTypeName method from String to QualifiedName. (#263)

• DBComposite's compositeTypeName method from String to QualifiedName. (#263)

• Changed Upsert by adding a predicate field, which allows partial indexes to be specified as conflict targets. (#264)

• The window functions lag, lead, firstValue, lastValue and nthValue can now operate on entire rows at once as opposed to just single columns. (#281)

Fixed

• Fixed a bug with catListTable and catNonEmptyTable where invalid SQL could be produced. (#240)

• A fix for #168, which prevented using catListTable on arrays of arrays. To achieve this we had to coerce arrays of arrays to text internally, which unfortunately isn't completely transparent; you can oberve it if you write something like listTable [listTable [10]] > listTable [listTable [9]]: previously that would be false, but now it's true. Arrays of non-arrays are unaffected by this.

• Fixes #228 where it was impossible to call nextval with a qualified sequence name.

• Fixes #71.

• Fixed a typo in the documentation for /=.. (#312)

• Fixed a bug where fromRational could crash with repeating fractions. (#309)

• Fixed a typo in the documentation for min. (#306)

1.4.1.0 (2023-01-19)

New features

• Rel8 now supports window functions. See the "Window functions" section of the Rel8 module documentation for more details. (#182)
• Query now has Monoid and Semigroup instances. (#207)
• createOrReplaceView has been added (to run CREATE OR REPLACE VIEW). (#209 and #212)
• deriving Rel8able now supports more polymorphism. (#215)
• Support GHC 9.4 (#199)

Bug fixes

• Insertion of DEFAULT values has been fixed. (#206)
• Avoid some exponential SQL generation in Rel8.Tabulate.alignWith. (#213)
• nextVal has been fixed to work with case-sensitive sequence names. (#217)

Other

• Correct the documentation for "Supplying Rel8able instances" (#200)
• Removed some redundant internal code (#202)
• Rel8 is now less dependant on the internal Opaleye API. (#204)

1.4.0.0 (2022-08-17)

Breaking changes

• The behavior of greatest/least has been corrected, and was previously flipped. (#183)

New features

• NullTable/HNull have been added. This is an alternative to MaybeTable that doesn't use a tag columns. It's less flexible (no Functor or Applicative instance) and is meaningless when used with a table that has no non-nullable columns (so nesting NullTable is redundant). But in situations where the underlying Table does have non-nullable columns, it can losslessly converted to and from MaybeTable. It is useful for embedding into a base table when you don't want to store the extra tag column in your schema. (#173)
• Add fromMaybeTable. (#179)
• Add alignMaybeTable. (#196)

Improvements

• Optimize implementation of AltTable for Tabulation (#178)

Other

• Documentation improvements for HADT. (#177)
• Document example usage of groupBy. (#184)
• Build with and require Opaleye >= 0.9.3.3. (#190)
• Build with hasql 1.6. (#195)

1.3.1.0 (2022-01-20)

Other

• Rel8 now requires Opaleye >= 0.9.1. (#165)

1.3.0.0 (2022-01-31)

Breaking changes

• div and mod have been changed to match Haskell semantics. If you need the PostgreSQL div() and mod() functions, use quot and rem. While this is not an API change, we feel this is a breaking change in semantics and have bumped the major version number. (#155)

New features

• divMod and quotRem functions have been added, matching Haskell's Prelude functions. (#155)
• avg and mode aggregation functions to find the mean value of an expression, or the most common row in a query, respectively. (#152)
• The full EqTable and OrdTable classes have been exported, allowing for instances to be manually created. (#157)
• Added like and ilike (for the LIKE and ILIKE operators). (#146)

Other

• Rel8 now requires Opaleye 0.9. (#158)
• Rel8's test suite supports Hedgehog 1.1. (#160)
• The documentation for binary operations has been corrected. (#162)

1.2.2.0 (2021-11-21)

Other

• Support GHC 9.2 (#145)
• Correct the documentation for except (#147)
• Support hasql 1.5 (#149)

1.2.1.0 (2021-11-21)

New features

• castTable has been added, which casts all columns selected from in Query according to the types of columns. This can occasionally be useful when using many/some on older PostgreSQL versions. (#137)

• Added <:, <=:, >:, >=:, greatest and least, which sort all columns in a table lexicographically. These operators are like the <. operators which operate on Exprs, but the <: operate on entire tables. (#139)

Other

• Support opaleye-0.8 (#142)

1.2.0.0 (2021-10-22)

New features

• New $* and $+ operators for projecting out of ListTable and NonEmptyTable respectively (analogous to the existing $? for MaybeTable). (#125)

Bug fixes

• Fix UPSERT support (and add tests), which was broken due to a typo inside Rel8 that made it impossible to construct Projections which are needed by UPSERT. (#134)

• Remove DBMin and DBMax instances for Bool, which are invalid as far as Postgres is concerned. and and or can be used to achieve the same thing. (#126)

• Fix aggregateMaybeTable, aggregateEitherTable and aggregateTheseTable, which generated invalid SQL previously. (#127)

Breaking changes

• rebind now takes an additional argument for the "name" of the binding. (#128)

1.1.0.0 (2021-07-16)

New features

• You can derive Rel8able for "vanilla" higher-kinded data types - data types that don't use the Column type family. For example, the following is now possible:

  data MyTable f = MyTable
    { myTableX :: f Int32
    , myTableY :: f Text
    } deriving (Generic, Rel8able)
  

• The Projectable and Biprojectable type classes have been introduced. These classes model a type of functors and bifunctors (respectively) on tables, where the mapping function is only able to rearrange columns. For example, the following is now possible:

  x <- project myTableX <$> many ...
  

• Insert's onConflict field now has a new option - DoUpdate. This maps to PostgreSQL's ON CONFLICT DO UPDATE syntax, which provides upsert support. (#85)

• Insert, Update and Delete has all been expanded to work with Querys. In particular:

  • Delete now supports a USING sub-query
  • Update now supports a FROM sub-query
  • Insert can delete the result of any arbitrary Query (providing the types match)

• The new Rel8.Tabulate module has been added, which gives a Map-esque interface to writing and composing queries. (#70)

• The new indexed Query -> Query function was added. This function augments each row in a query with it's 0-based index. (#117)

Breaking changes

• Insert, Delete and Update have all changed. In particular, for Insert users should now replace rows = xs with rows = values xs. (#85)

• Rename whereExists and whereNotExists to present and absent respectively. (#57)

• Simplify evaluate to run directly inside the Query monad, rendering the Evaluate monad unnecessary. (#56)

• The Labelable type class has been removed as it's no longer necessary. This class was mostly internal, though was visible in some public API functions. (#76)

• EitherTable, ListTable, MaybeTable, NonEmptyTable and TheseTable now all have an extra parameter which tracks the context of the table. If you manually specified any types before, you need to provide this parameter - usually Expr will be the correct choice. (#101)

• We've done a lot of work to simplify the internals of Rel8, but some of these are breaking changes. However, most of these API changes shouldn't affect users as they are mostly types used in type inference, or affect classes where users don't need to define their own instances.

  • The kind of a Rel8able is now (Type -> Type) -> Type. Previously it was (X -> Type) -> Type.
  • Table has a new associated type - FromExprs. This was previously an open type family.
  • Table has a new associated type - Transpose - and Recontextualise has been renamed to Transposes. This Transposes class now operates in terms of Transpose.

• select, insert, update and delete now produce Hasql Statements, rather than actually running the statement as IO. This allows Rel8 to be used with transaction/connection-managing monads like hasql-transaction. (#94)

Bug fixes

• Fixes a bug where cartesian products of queries using catListTable, catNonEmptyTable, catList and catNonEmpty would incorrectly be zipped instead. (#61)

• Require Opaleye 0.7.3.0. This version has better support for lateral queries, which can improve query plans, especially in optional/LEFT JOIN (#72)

• Rel8 now compiles with GHC 9. (#97)

Other changes

• exists is now implemented in terms of the SQL EXISTS keyword. (#69)

• alignBy no longer requires Tables. (#67)

1.0.0.1 (2021-06-21)

This release contains various fixes for documentation.

1.0.0.0 (2021-06-18)

• Initial release.