Database.Esqueleto.Legacy

WHERE clause: restrict the query's result.

An ON clause, useful to describe how two tables are related. Cross joins and tuple-joins do not need an on clause, but InnerJoin and the various outer joins do.

Database.Esqueleto.Experimental in version 4.0.0.0 of the library. The Experimental module has a dramatically improved means for introducing tables and entities that provides more power and less potential for runtime errors.

If you don't include an on clause (or include too many!) then a runtime exception will be thrown.

As an example, consider this simple join:

select $
from $ \(foo `InnerJoin` bar) -> do
  on (foo ^. FooId ==. bar ^. BarFooId)
  ...

We need to specify the clause for joining the two columns together. If we had this:

select $
from $ \(foo `CrossJoin` bar) -> do
  ...

Then we can safely omit the on clause, because the cross join will make pairs of all records possible.

You can do multiple on clauses in a query. This query joins three tables, and has two on clauses:

select $
from $ \(foo `InnerJoin` bar `InnerJoin` baz) -> do
  on (baz ^. BazId ==. bar ^. BarBazId)
  on (foo ^. FooId ==. bar ^. BarFooId)
  ...

Old versions of esqueleto required that you provide the on clauses in reverse order. This restriction has been lifted - you can now provide on clauses in any order, and the SQL should work itself out. The above query is now totally equivalent to this:

select $
from $ \(foo `InnerJoin` bar `InnerJoin` baz) -> do
  on (foo ^. FooId ==. bar ^. BarFooId)
  on (baz ^. BazId ==. bar ^. BarBazId)
  ...

GROUP BY clause. You can enclose multiple columns in a tuple.

select $ from \(foo `InnerJoin` bar) -> do
  on (foo ^. FooBarId ==. bar ^. BarId)
  groupBy (bar ^. BarId, bar ^. BarName)
  return (bar ^. BarId, bar ^. BarName, countRows)

With groupBy you can sort by aggregate functions, like so (we used let to restrict the more general countRows to SqlSqlExpr (Value Int) to avoid ambiguity---the second use of countRows has its type restricted by the :: Int below):

r <- select $ from \(foo `InnerJoin` bar) -> do
  on (foo ^. FooBarId ==. bar ^. BarId)
  groupBy $ bar ^. BarName
  let countRows' = countRows
  orderBy [asc countRows']
  return (bar ^. BarName, countRows')
forM_ r $ \(Value name, Value count) -> do
  print name
  print (count :: Int)

Need more columns?

The ToSomeValues class is defined for SqlExpr and tuples of SqlExprs. We only have definitions for up to 8 elements in a tuple right now, so it's possible that you may need to have more than 8 elements.

For example, consider a query with a groupBy call like this:

groupBy (e0, e1, e2, e3, e4, e5, e6, e7)

This is the biggest you can get with a single tuple. However, you can easily nest the tuples to add more:

groupBy ((e0, e1, e2, e3, e4, e5, e6, e7), e8, e9)

ORDER BY clause. See also asc and desc.

Multiple calls to orderBy get concatenated on the final query, including distinctOnOrderBy.

Ascending order of this field or SqlExpression.

Descending order of this field or SqlExpression.

LIMIT. Limit the number of returned rows.

OFFSET. Usually used with limit.

DISTINCT. Change the current SELECT into SELECT DISTINCT. For example:

select $ distinct $
  from \foo -> do
  ...

Note that this also has the same effect:

select $
  from \foo -> do
  distinct (return ())
  ...

DISTINCT ON. Change the current SELECT into SELECT DISTINCT ON (SqlExpressions). For example:

select $
  from \foo ->
  distinctOn [don (foo ^. FooName), don (foo ^. FooState)] $ do
  ...

You can also chain different calls to distinctOn. The above is equivalent to:

select $
  from \foo ->
  distinctOn [don (foo ^. FooName)] $
  distinctOn [don (foo ^. FooState)] $ do
  ...

Each call to distinctOn adds more SqlExpressions. Calls to distinctOn override any calls to distinct.

Note that PostgreSQL requires the SqlExpressions on DISTINCT ON to be the first ones to appear on a ORDER BY. This is not managed automatically by esqueleto, keeping its spirit of trying to be close to raw SQL.

Supported by PostgreSQL only.

Erase an SqlExpression's type so that it's suitable to be used by distinctOn.

A convenience function that calls both distinctOn and orderBy. In other words,

distinctOnOrderBy [asc foo, desc bar, desc quux] $ do
  ...

is the same as:

distinctOn [don foo, don  bar, don  quux] $ do
  orderBy  [asc foo, desc bar, desc quux]
  ...

HAVING.

Add a locking clause to the query. Please read LockingKind documentation and your RDBMS manual. Unsafe since not all locking clauses are implemented for every RDBMS

If multiple calls to locking are made on the same query, the last one is used.

Project a field of an entity.

Project an EntityField of a nullable entity. The result type will be Nullable, meaning that nested Maybe won't be produced here.

As of v3.6.0.0, this will attempt to combine nested Maybe. If you want to keep nested Maybe, then see ??..

Lift a constant value from Haskell-land to the query.

IS NULL comparison.

For IS NOT NULL, you can negate this with not_, as in not_ (isNothing (person ^. PersonAge))

Warning: Persistent and Esqueleto have different behavior for != Nothing:

In SQL, = NULL and != NULL return NULL instead of true or false. For this reason, you very likely do not want to use !=. Nothing in Esqueleto. You may find these hlint rules helpful to enforce this:

- error: {lhs: v Database.Esqueleto.==. Database.Esqueleto.nothing, rhs: Database.Esqueleto.isNothing v, name: Use Esqueleto's isNothing}
- error: {lhs: v Database.Esqueleto.==. Database.Esqueleto.val Nothing, rhs: Database.Esqueleto.isNothing v, name: Use Esqueleto's isNothing}
- error: {lhs: v Database.Esqueleto.!=. Database.Esqueleto.nothing, rhs: not_ (Database.Esqueleto.isNothing v), name: Use Esqueleto's not isNothing}
- error: {lhs: v Database.Esqueleto.!=. Database.Esqueleto.val Nothing, rhs: not_ (Database.Esqueleto.isNothing v), name: Use Esqueleto's not isNothing}

Analogous to Just, promotes a value of type typ into one of type Maybe typ. It should hold that val . Just === just . val.

This function will try not to produce a nested Maybe. This is in accord with how SQL represents NULL. That means that just . just = just. This behavior was changed in v3.6.0.0. If you want to produce nested Maybe, see just'.

Like just, but this function does not try to collapse nested Maybe. This may be useful if you have type inference problems with just.

NULL value.

Join nested Maybes in a Value into one. This is useful when calling aggregate functions on nullable fields.

As of v3.6.0.0, this function will attempt to work on both SqlExpr (Value (Maybe a)) as well as SqlExpr (Value (Maybe (Maybe a))) inputs to make transitioning to NullableFieldProjection easier. This may make type inference worse in some cases. If you want the monomorphic variant, see joinV'

Like joinV, but monomorphic: the input type only works on SqlExpr (Value (Maybe (Maybe a))).

This function may be useful if you have type inference issues with joinV.

Project an SqlExpression that may be null, guarding against null cases.

COUNT(*) value.

COUNT.

COUNT(DISTINCT x).

This operator produces the SQL operator =, which is used to compare values for equality.

Example:

 query :: UserId -> SqlPersistT IO [Entity User]
 query userId = select $ do
     user <- from $ table @User
     where_ (user ^. UserId ==. val userId)
     pure user

This would generate the following SQL:

 SELECT user.*
 FROM user
 WHERE user.id = ?

This operator translates to the SQL operator >=.

Example:

 where_ $ user ^. UserAge >=. val 21

This operator translates to the SQL operator >.

Example:

 where_ $ user ^. UserAge >. val 20

This operator translates to the SQL operator <=.

Example:

 where_ $ val 21 <=. user ^. UserAge

This operator translates to the SQL operator <.

Example:

 where_ $ val 20 <. user ^. UserAge

This operator translates to the SQL operator !=.

Example:

 where_ $ user ^. UserName !=. val Bob

This operator translates to the SQL operator AND.

Example:

 where_ $
         user ^. UserName ==. val Matt
     &&. user ^. UserAge >=. val 21

This operator translates to the SQL operator AND.

Example:

 where_ $
         user ^. UserName ==. val Matt
     ||. user ^. UserName ==. val John

a between (b, c) translates to the SQL expression a >= b AND a <= c. It does not use a SQL BETWEEN operator.

@since: 3.1.0

This operator translates to the SQL operator +.

This does not require or assume anything about the SQL values. Interpreting what +. means for a given type is left to the database engine.

Example:

 user ^. UserAge +. val 10

This operator translates to the SQL operator -.

This does not require or assume anything about the SQL values. Interpreting what -. means for a given type is left to the database engine.

Example:

 user ^. UserAge -. val 10

This operator translates to the SQL operator /.

This does not require or assume anything about the SQL values. Interpreting what /. means for a given type is left to the database engine.

Example:

 user ^. UserAge /. val 10

This operator translates to the SQL operator *.

This does not require or assume anything about the SQL values. Interpreting what *. means for a given type is left to the database engine.

Example:

 user ^. UserAge *. val 10

Allow a number of one type to be used as one of another type via an implicit cast. An explicit cast is not made, this function changes only the types on the Haskell side.

Caveat: Trying to use castNum from Double to Int will not result in an integer, the original fractional number will still be used! Use round_, ceiling_ or floor_ instead.

Safety: This operation is mostly safe due to the Num constraint between the types and the fact that RDBMSs usually allow numbers of different types to be used interchangeably. However, there may still be issues with the query not being accepted by the RDBMS or persistent not being able to parse it.

Same as castNum, but for nullable values.

COALESCE function. Evaluates the arguments in order and returns the value of the first non-NULL SqlExpression, or NULL (Nothing) otherwise. Some RDBMSs (such as SQLite) require at least two arguments; please refer to the appropriate documentation.

Like coalesce, but takes a non-nullable SqlExpression placed at the end of the SqlExpression list, which guarantees a non-NULL result.

LOWER function.

UPPER function. @since 3.3.0

TRIM function. @since 3.3.0

LTRIM function. @since 3.3.0

RTRIM function. @since 3.3.0

LENGTH function. @since 3.3.0

LEFT function. @since 3.3.0

RIGHT function. @since 3.3.0

LIKE operator.

ILIKE operator (case-insensitive LIKE).

Supported by PostgreSQL only. Deprecated in version 3.6.0 in favor of the version available from Database.Esqueleto.PostgreSQL.

The string %. May be useful while using like and concatenation (concat_ or ++., depending on your database). Note that you always have to type the parenthesis, for example:

name `like` (%) ++. val "John" ++. (%)

The CONCAT function with a variable number of parameters. Supported by MySQL and PostgreSQL. SQLite supports this in versions after 3.44.0, and persistent-sqlite supports this in versions 2.13.3.0 and after.

The || string concatenation operator (named after Haskell's ++ in order to avoid naming clash with ||.).

Supported by SQLite and PostgreSQL.

MySQL support requires setting the SQL mode to PIPES_AS_CONCAT or ANSI - see this StackOverflow answer.

Cast a string type into Text. This function is very useful if you want to use newtypes, or if you want to apply functions such as like to strings of different types.

Safety: This is a slightly unsafe function, especially if you have defined your own instances of SqlString. Also, since Maybe is an instance of SqlString, it's possible to turn a nullable value into a non-nullable one. Avoid using this function if possible.

Execute a subquery SELECT in an SqlExpression. Returns a list of values.

Lift a list of constant value from Haskell-land to the query.

Same as just but for ValueList. Most of the time you won't need it, though, because you can use just from inside subList_select or Just from inside valList.

IN operator. For example if you want to select all Persons by a list of IDs:

SELECT *
FROM Person
WHERE Person.id IN (?)

In esqueleto, we may write the same query above as:

select $
from $ \person -> do
where_ $ person ^. PersonId `in_` valList personIds
return person

Where personIds is of type [Key Person].

NOT IN operator.

EXISTS operator. For example:

select $
from $ \person -> do
where_ $ exists $
         from $ \post -> do
         where_ (post ^. BlogPostAuthorId ==. person ^. PersonId)
return person

NOT EXISTS operator.

SET clause used on UPDATEs. Note that while it's not a type error to use this function on a SELECT, it will most certainly result in a runtime error.

CASE statement. For example:

select $
return $
case_
   [ when_
       (exists $
       from $ \p -> do
       where_ (p ^. PersonName ==. val "Mike"))
     then_
       (subSelect $
       from $ \v -> do
       let sub =
               from $ \c -> do
               where_ (c ^. PersonName ==. val "Mike")
               return (c ^. PersonFavNum)
       where_ (just (v ^. PersonFavNum) >. subSelect sub)
       return $ count (v ^. PersonName) +. val (1 :: Int)) ]
   (else_ $ val (-1))

This query is a bit complicated, but basically it checks if a person named "Mike" exists, and if that person does, run the subquery to find out how many people have a ranking (by Fav Num) higher than "Mike".

NOTE: There are a few things to be aware about this statement.

• This only implements the full CASE statement, it does not implement the "simple" CASE statement.

• At least one when_ and then_ is mandatory otherwise it will emit an error.

• The else_ is also mandatory, unlike the SQL statement in which if the ELSE is omitted it will return a NULL. You can reproduce this via nothing.

Convert an entity's key into another entity's.

This function is to be used when you change an entity's Id to be that of another entity. For example:

Bar
  barNum Int
Foo
  bar BarId
  fooNum Int
  Primary bar

In this example, Bar is said to be the BaseEnt(ity), and Foo the child. To model this in Esqueleto, declare:

instance ToBaseId Foo where
  type BaseEnt Foo = Bar
  toBaseIdWitness barId = FooKey barId

Now you're able to write queries such as:

select $
from $ (bar `InnerJoin` foo) -> do
on (toBaseId (foo ^. FooId) ==. bar ^. BarId)
return (bar, foo)

Note: this function may be unsafe to use in conditions not like the one of the example above.

The inverse of toBaseId. Note that this is somewhat less "safe" than toBaseId. Calling toBaseId will usually mean that a foreign key constraint is present that guarantees the presence of the base ID. fromBaseId has no such guarantee. Consider the code example given in toBaseId:

Bar
  barNum Int
Foo
  bar BarId
  fooNum Int
  Primary bar

instance ToBaseId Foo where
  type BaseEnt Foo = Bar
  toBaseIdWitness barId = FooKey barId

The type of toBaseId for Foo would be:

toBaseId :: SqlExpr (Value FooId) -> SqlExpr (Value BarId)

The foreign key constraint on Foo means that every FooId points to a BarId in the database. However, fromBaseId will not have this:

fromBaseId :: SqlExpr (Value BarId) -> SqlExpr (Value FooId)

As fromBaseId, but works on Maybe keys.

Like toBaseId, but works on Maybe keys.

Execute a subquery SELECT in a SqlExpr. The query passed to this function will only return a single result - it has a LIMIT 1 passed in to the query to make it safe, and the return type is Maybe to indicate that the subquery might result in 0 rows.

If you find yourself writing joinV . subSelect, then consider using subSelectMaybe.

If you're performing a countRows, then you can use subSelectCount which is safe.

If you know that the subquery will always return exactly one row (eg a foreign key constraint guarantees that you'll get exactly one row), then consider subSelectUnsafe, along with a comment explaining why it is safe.

Execute a subquery SELECT in a SqlExpr. This function is a shorthand for the common joinV . subSelect idiom, where you are calling subSelect on an expression that would be Maybe already.

As an example, you would use this function when calling sum_ or max_, which have Maybe in the result type (for a 0 row query).

Performs a COUNT of the given query in a subSelect manner. This is always guaranteed to return a result value, and is completely safe.

Performs a sub-select using the given foreign key on the entity. This is useful to extract values that are known to be present by the database schema.

As an example, consider the following persistent definition:

User
  profile ProfileId

Profile
  name    Text

The following query will return the name of the user.

getUserWithName =
    select $
    from $ user ->
    pure (user, subSelectForeign user UserProfile (^. ProfileName)

Execute a subquery SELECT in a SqlExpr that returns a list. This is an alias for subList_select and is provided for symmetry with the other safe subselect functions.

Execute a subquery SELECT in a SqlExpr. This function is unsafe, because it can throw runtime exceptions in two cases:

1. If the query passed has 0 result rows, then it will return a NULL value. The persistent parsing operations will fail on an unexpected NULL.

2. If the query passed returns more than one row, then the SQL engine will fail with an error like "More than one row returned by a subquery used as an expression".

This function is safe if you guarantee that exactly one row will be returned, or if the result already has a Maybe type for some reason.

For variants with the safety encoded already, see subSelect and subSelectMaybe. For the most common safe use of this, see subSelectCount.

Class that enables one to use toBaseId to convert an entity's key on a query into another (cf. toBaseId).

Syntax sugar for case_.

Syntax sugar for case_.

Syntax sugar for case_.

FROM clause: bring entities into scope.

Note that this function will be replaced by the one in Database.Esqueleto.Experimental in version 4.0.0.0 of the library. The Experimental module has a dramatically improved means for introducing tables and entities that provides more power and less potential for runtime errors.

This function internally uses two type classes in order to provide some flexibility of how you may call it. Internally we refer to these type classes as the two different magics.

The innermost magic allows you to use from with the following types:

• expr (Entity val), which brings a single entity into scope.

• expr (Maybe (Entity val)), which brings a single entity that may be NULL into scope. Used for OUTER JOINs.

• A JOIN of any other two types allowed by the innermost magic, where a JOIN may be an InnerJoin, a CrossJoin, a LeftOuterJoin, a RightOuterJoin, or a FullOuterJoin. The JOINs have left fixity.

The outermost magic allows you to use from on any tuples of types supported by innermost magic (and also tuples of tuples, and so on), up to 8-tuples.

Note that using from for the same entity twice does work and corresponds to a self-join. You don't even need to use two different calls to from, you may use a JOIN or a tuple.

The following are valid examples of uses of from (the types of the arguments of the lambda are inside square brackets):

from $ \person -> ...
from $ \(person, blogPost) -> ...
from $ \(p `LeftOuterJoin` mb) -> ...
from $ \(p1 `InnerJoin` f `InnerJoin` p2) -> ...
from $ \((p1 `InnerJoin` f) `InnerJoin` p2) -> ...

The types of the arguments to the lambdas above are, respectively:

person
  :: ( Esqueleto query expr backend
     , PersistEntity Person
     , PersistEntityBackend Person ~ backend
     ) => expr (Entity Person)
(person, blogPost)
  :: (...) => (expr (Entity Person), expr (Entity BlogPost))
(p `LeftOuterJoin` mb)
  :: (...) => InnerJoin (expr (Entity Person)) (expr (Maybe (Entity BlogPost)))
(p1 `InnerJoin` f `InnerJoin` p2)
  :: (...) => InnerJoin
                (InnerJoin (expr (Entity Person))
                           (expr (Entity Follow)))
                (expr (Entity Person))
(p1 `InnerJoin` (f `InnerJoin` p2)) ::
  :: (...) => InnerJoin
                (expr (Entity Person))
                (InnerJoin (expr (Entity Follow))
                           (expr (Entity Person)))

Note that some backends may not support all kinds of JOINs.

A single value (as opposed to a whole entity). You may use (^.) or (?.) to get a Value from an Entity.

A list of single values. There's a limited set of functions able to work with this data type (such as subList_select, valList, in_ and exists).

Phantom type used by orderBy, asc and desc.

Phantom type used by distinctOn and don.

Different kinds of locking clauses supported by locking.

Note that each RDBMS has different locking support. The constructors of this datatype specify only the syntax of the locking mechanism, not its semantics. For example, even though both MySQL and PostgreSQL support ForUpdate, there are no guarantees that they will behave the same.

FOR UPDATE syntax.

Usage:

 locking forUpdate

FOR UPDATE SKIP LOCKED syntax.

 locking forUpdateSkipLocked

Lockable entity

Example use:

select $ do
    (p :& bp) <- from $
        table Person
        innerJoin table BlogPost
            on do
                (p :& bp) -> p ^. PersonId ==. b ^. BlogPostAuthorId
    forUpdateOf (p :& b) skipLocked
    return p

Phantom class of data types that are treated as strings by the RDBMS. It has no methods because it's only used to avoid type errors such as trying to concatenate integers.

If you have a custom data type or newtype, feel free to make it an instance of this class.

SQL backend for esqueleto using SqlPersistT.

An expression on the SQL backend.

Raw expression: Contains a SqlExprMeta and a function for building the expr. It recieves a parameter telling it whether it is in a parenthesized context, and takes information about the SQL connection (mainly for escaping names) and returns both an string (Builder) and a list of values to be interpolated by the SQL backend.

Constraint synonym for persistent entities whose backend is SqlBackend.

Execute an esqueleto SELECT query inside persistent's SqlPersistT monad and return a list of rows.

We've seen that from has some magic about which kinds of things you may bring into scope. This select function also has some magic for which kinds of things you may bring back to Haskell-land by using SqlQuery's return:

• You may return a SqlExpr (Entity v) for an entity v (i.e., like the * in SQL), which is then returned to Haskell-land as just Entity v.

• You may return a SqlExpr (Maybe (Entity v)) for an entity v that may be NULL, which is then returned to Haskell-land as Maybe (Entity v). Used for OUTER JOINs.

• You may return a SqlExpr (Value t) for a value t (i.e., a single column), where t is any instance of PersistField, which is then returned to Haskell-land as Value t. You may use Value to return projections of an Entity (see (^.) and (?.)) or to return any other value calculated on the query (e.g., countRows or subSelect).

The SqlSelect a r class has functional dependencies that allow type information to flow both from a to r and vice-versa. This means that you'll almost never have to give any type signatures for esqueleto queries. For example, the query select $ from $ \p -> return p alone is ambiguous, but in the context of

do ps <- select $
         from $ \p ->
         return p
   liftIO $ mapM_ (putStrLn . personName . entityVal) ps

we are able to infer from that single personName . entityVal function composition that the p inside the query is of type SqlExpr (Entity Person).

Execute an esqueleto SELECT query inside persistent's SqlPersistT monad and return the first entry wrapped in a Maybe. @since 3.5.1.0

Example usage

firstPerson :: MonadIO m => SqlPersistT m (Maybe (Entity Person))
firstPerson =
 selectOne $ do
     person <- from $ table @Person
     return person

The above query is equivalent to a select combined with limit but you would still have to transform the results from a list:

firstPerson :: MonadIO m => SqlPersistT m [Entity Person]
firstPerson =
 select $ do
     person <- from $ table @Person
     limit 1
     return person

Execute an esqueleto SELECT query inside persistent's SqlPersistT monad and return a Source of rows.

Execute an esqueleto DELETE query inside persistent's SqlPersistT monad. Note that currently there are no type checks for statements that should not appear on a DELETE query.

Example of usage:

delete $
from $ \appointment ->
where_ (appointment ^. AppointmentDate <. val now)

Unlike select, there is a useful way of using delete that will lead to type ambiguities. If you want to delete all rows (i.e., no where_ clause), you'll have to use a type signature:

delete $
from $ \(appointment :: SqlExpr (Entity Appointment)) ->
return ()

Database.Esqueleto.Experimental:

 delete $ do
   userFeature <- from $ table @UserFeature
   where_ ((userFeature ^. UserFeatureFeature) notIn valList allKnownFeatureFlags)

Same as delete, but returns the number of rows affected.

Execute an esqueleto UPDATE query inside persistent's SqlPersistT monad. Note that currently there are no type checks for statements that should not appear on a UPDATE query.

Example of usage:

update $ \p -> do
set p [ PersonAge =. just (val thisYear) -. p ^. PersonBorn ]
where_ $ isNothing (p ^. PersonAge)

Same as update, but returns the number of rows affected.

Insert a PersistField for every selected value.

Insert a PersistField for every selected value, return the count afterward

Apply a PersistField constructor to SqlExpr Value arguments.

Apply extra SqlExpr Value arguments to a PersistField constructor

(Internal) Class that implements the tuple from magic (see fromStart).

valkey i = val . toSqlKey (https://github.com/prowdsponsor/esqueleto/issues/9).

valJ is like val but for something that is already a Value. The use case it was written for was, given a Value lift the Key for that Value into the query expression in a type safe way. However, the implementation is more generic than that so we call it valJ.

Its important to note that the input entity and the output entity are constrained to be the same by the type signature on the function (https://github.com/prowdsponsor/esqueleto/pull/69).

Avoid N+1 queries and join entities into a map structure.

This function is useful to call on the result of a single JOIN. For example, suppose you have this query:

getFoosAndNestedBarsFromParent
    :: ParentId
    -> SqlPersistT IO [(Entity Foo, Maybe (Entity Bar))]
getFoosAndNestedBarsFromParent parentId =
    select $ do
        (foo :& bar) <- from $
            table Foo
            `LeftOuterJoin`
            table Bar
                `on` do
                    \(foo :& bar) ->
                        foo ^. FooId ==. bar ?. BarFooId
        where_ $
            foo ^. FooParentId ==. val parentId
        pure (foo, bar)

This is a natural result type for SQL - a list of tuples. However, it's not what we usually want in Haskell - each Foo in the list will be represented multiple times, once for each Bar.

We can write fmap associateJoin and it will translate it into a Map that is keyed on the Key of the left Entity, and the value is a tuple of the entity's value as well as the list of each coresponding entity.

getFoosAndNestedBarsFromParentHaskellese
    :: ParentId
    -> SqlPersistT (Map (Key Foo) (Foo, [Maybe (Entity Bar)]))
getFoosAndNestedBarsFromParentHaskellese parentId =
    fmap associateJoin $ getFoosdAndNestedBarsFromParent parentId

What if you have multiple joins?

Let's use associateJoin with a *two* join query.

userPostComments
    :: SqlQuery (SqlExpr (Entity User, Entity Post, Entity Comment))
userPostsComment = do
    (u :& p :& c) <- from $
        table User
        `InnerJoin`
        table Post
            on do
                \(u :& p) ->
                    u ^. UserId ==. p ^. PostUserId
        `InnerJoin`
        table @Comment
            `on` do
                \(_ :& p :& c) ->
                    p ^. PostId ==. c ^. CommentPostId
    pure (u, p, c)

This query returns a User, with all of the users Posts, and then all of the Comments on that post.

First, we *nest* the tuple.

nest :: (a, b, c) -> (a, (b, c))
nest (a, b, c) = (a, (b, c))

This makes the return of the query conform to the input expected from associateJoin.

nestedUserPostComments
    :: SqlPersistT IO [(Entity User, (Entity Post, Entity Comment))]
nestedUserPostComments =
    fmap nest $ select userPostsComments

Now, we can call associateJoin on it.

associateUsers
    :: [(Entity User, (Entity Post, Entity Comment))]
    -> Map UserId (User, [(Entity Post, Entity Comment)])
associateUsers =
    associateJoin

Next, we'll use the Functor instances for Map and tuple to call associateJoin on the [(Entity Post, Entity Comment)].

associatePostsAndComments
    :: Map UserId (User, [(Entity Post, Entity Comment)])
    -> Map UserId (User, Map PostId (Post, [Entity Comment]))
associatePostsAndComments =
    fmap (fmap associateJoin)

For more reading on this topic, see this Foxhound Systems blog post.

Synonym for delete that does not clash with esqueleto's delete.

A more general way to convert instances of ToJSON type class to strict text Text.

Predefined parseJSON. The input JSON looks like {"id": 1, "name": ...}.

The typical usage is:

instance FromJSON (Entity User) where
    parseJSON = entityIdFromJSON

Predefined toJSON. The resulting JSON looks like {"id": 1, "name": ...}.

The typical usage is:

instance ToJSON (Entity User) where
    toJSON = entityIdToJSON

Get list of values corresponding to given entity.

Convenience function for getting a free PersistField instance from a type with JSON instances. The JSON parser used will accept JSON values other that object and arrays. So, if your instance serializes the data to a JSON string, this will still work.

Example usage in combination with toPersistValueJSON:

instance PersistField MyData where
  fromPersistValue = fromPersistValueJSON
  toPersistValue = toPersistValueJSON

Predefined parseJSON. The input JSON looks like {"key": 1, "value": {"name": ...}}.

The typical usage is:

instance FromJSON (Entity User) where
    parseJSON = keyValueEntityFromJSON

Predefined toJSON. The resulting JSON looks like {"key": 1, "value": {"name": ...}}.

The typical usage is:

instance ToJSON (Entity User) where
    toJSON = keyValueEntityToJSON

Construct an Entity record by providing a value for each of the record's fields.

These constructions are equivalent:

entityMattConstructor, entityMattTabulate :: Entity User
entityMattConstructor =
    Entity
        { entityKey = toSqlKey 123
        , entityVal =
            User
                { userName = Matt
                , userAge = 33
                }
        }

entityMattTabulate =
    tabulateEntity $ \case
        UserId ->
            toSqlKey 123
        UserName ->
            Matt
        UserAge ->
            33

This is a specialization of tabulateEntityA, which allows you to construct an Entity by providing an Applicative action for each field instead of a regular function.

Convenience function for getting a free PersistField instance from a type with JSON instances.

Example usage in combination with fromPersistValueJSON:

instance PersistField MyData where
  fromPersistValue = fromPersistValueJSON
  toPersistValue = toPersistValueJSON

Get the Keys of all records matching the given criterion.

For an example, see selectList.

Curry this to make a convenience function that loads an associated model.

foreign = belongsTo foreignId

Same as belongsTo, but uses getJust and therefore is similarly unsafe.

Like get, but returns the complete Entity.

Example usage

With schema-1 and dataset-1,

getSpjEntity :: MonadIO m => ReaderT SqlBackend m (Maybe (Entity User))
getSpjEntity = getEntity spjId

mSpjEnt <- getSpjEntity

The above query when applied on dataset-1, will get this entity:

+----+------+-----+
| id | name | age |
+----+------+-----+
|  1 | SPJ  |  40 |
+----+------+-----+

Same as get, but for a non-null (not Maybe) foreign key. Unsafe unless your database is enforcing that the foreign key is valid.

Example usage

With schema-1 and dataset-1,

getJustSpj :: MonadIO m => ReaderT SqlBackend m User
getJustSpj = getJust spjId

spj <- getJust spjId

The above query when applied on dataset-1, will get this record:

+----+------+-----+
| id | name | age |
+----+------+-----+
|  1 | SPJ  |  40 |
+----+------+-----+

getJustUnknown :: MonadIO m => ReaderT SqlBackend m User
getJustUnknown = getJust unknownId

mrx <- getJustUnknown

This just throws an error.

Same as getJust, but returns an Entity instead of just the record.

Example usage

With schema-1 and dataset-1,

getJustEntitySpj :: MonadIO m => ReaderT SqlBackend m (Entity User)
getJustEntitySpj = getJustEntity spjId

spjEnt <- getJustEntitySpj

The above query when applied on dataset-1, will get this entity:

+----+------+-----+
| id | name | age |
+----+------+-----+
|  1 | SPJ  |  40 |
+----+------+-----+

Like insert, but returns the complete Entity.

Example usage

With schema-1 and dataset-1,

insertHaskellEntity :: MonadIO m => ReaderT SqlBackend m (Entity User)
insertHaskellEntity = insertEntity $ User "Haskell" 81

haskellEnt <- insertHaskellEntity

The above query when applied on dataset-1, will produce this:

+----+---------+-----+
| id |  name   | age |
+----+---------+-----+
|  1 | SPJ     |  40 |
+----+---------+-----+
|  2 | Simon   |  41 |
+----+---------+-----+
|  3 | Haskell |  81 |
+----+---------+-----+

Like insertEntity but just returns the record instead of Entity.

Example usage

With schema-1 and dataset-1,

insertDaveRecord :: MonadIO m => ReaderT SqlBackend m User
insertDaveRecord = insertRecord $ User "Dave" 50

dave <- insertDaveRecord

The above query when applied on dataset-1, will produce this:

+-----+------+-----+
|id   |name  |age  |
+-----+------+-----+
|1    |SPJ   |40   |
+-----+------+-----+
|2    |Simon |41   |
+-----+------+-----+
|3    |Dave  |50   |
+-----+------+-----+

Run a query against a larger backend by plucking out BaseBackend backend

This is a helper for reusing existing queries when expanding the backend type.

Run a query against a compatible backend, by projecting the backend

This is a helper for using queries which run against a specific backend type that your backend is compatible with.

Check whether there are any conflicts for unique keys with this entity and existing entities in the database.

Returns Nothing if the entity would be unique, and could thus safely be inserted. on a conflict returns the conflicting key

Example usage

We use schema-1 and dataset-1 here.

This would be Nothing:

mAlanConst <- checkUnique $ User "Alan" 70

While this would be Just because SPJ already exists:

mSpjConst <- checkUnique $ User "SPJ" 60

Check whether there are any conflicts for unique keys with this entity and existing entities in the database.

Returns Nothing if the entity would stay unique, and could thus safely be updated. on a conflict returns the conflicting key

This is similar to checkUnique, except it's useful for updating - when the particular entity already exists, it would normally conflict with itself. This variant ignores those conflicts

Example usage

We use schema-1 and dataset-1 here.

This would be Nothing:

mAlanConst <- checkUnique $ User "Alan" 70

While this would be Just because SPJ already exists:

mSpjConst <- checkUnique $ User "SPJ" 60

A modification of getBy, which takes the PersistEntity itself instead of a Unique record. Returns a record matching one of the unique keys. This function makes the most sense on entities with a single Unique constructor.

Example usage

With schema-1 and dataset-1,

getBySpjValue :: MonadIO m => ReaderT SqlBackend m (Maybe (Entity User)) getBySpjValue = getByValue $ User SPJ 999

mSpjEnt <- getBySpjValue

The above query when applied on dataset-1, will get this record:

+----+------+-----+
| id | name | age |
+----+------+-----+
|  1 | SPJ  |  40 |
+----+------+-----+

Insert a value, checking for conflicts with any unique constraints. If a duplicate exists in the database, it is returned as Left. Otherwise, the new 'Key is returned as Right.

Example usage

With schema-2 and dataset-1, we have following lines of code:

l1 <- insertBy $ User "SPJ" 20
l2 <- insertBy $ User "XXX" 41
l3 <- insertBy $ User "SPJ" 40
r1 <- insertBy $ User "XXX" 100

First three lines return Left because there're duplicates in given record's uniqueness constraints. While the last line returns a new key as Right.

Like insertEntity, but returns Nothing when the record couldn't be inserted because of a uniqueness constraint.

Example usage

We use schema-2 and dataset-1 here.

insertUniqueSpjEntity :: MonadIO m => ReaderT SqlBackend m (Maybe (Entity User))
insertUniqueSpjEntity = insertUniqueEntity $ User "SPJ" 50

mSpjEnt <- insertUniqueSpjEntity

The above query results Nothing as SPJ already exists.

insertUniqueAlexaEntity :: MonadIO m => ReaderT SqlBackend m (Maybe (Entity User))
insertUniqueAlexaEntity = insertUniqueEntity $ User "Alexa" 3

mAlexaEnt <- insertUniqueSpjEntity

Because there's no such unique keywords of the given record, the above query when applied on dataset-1, will produce this:

+----+-------+-----+
| id | name  | age |
+----+-------+-----+
|  1 | SPJ   |  40 |
+----+-------+-----+
|  2 | Simon |  41 |
+----+-------+-----+
|  3 | Alexa |   3 |
+----+-------+-----+

Given a proxy for a PersistEntity record, this returns the sole UniqueDef for that entity.

Return the single unique key for a record.

Example usage

We use shcema-1 and dataset-1 here.

onlySimonConst :: MonadIO m => ReaderT SqlBackend m (Unique User)
onlySimonConst = onlyUnique $ User "Simon" 999

mSimonConst <- onlySimonConst

mSimonConst would be Simon's uniqueness constraint. Note that onlyUnique doesn't work if there're more than two constraints. It will fail with a type error instead.

Attempt to replace the record of the given key with the given new record. First query the unique fields to make sure the replacement maintains uniqueness constraints.

Return Nothing if the replacement was made. If uniqueness is violated, return a Just with the Unique violation

Return the database name for the given entity.

Retrieve the list of FieldDef that makes up the fields of the entity.

This does not return the fields for an Id column or an implicit id. It will return the key columns if you used the Primary syntax for defining the primary key.

This does not return fields that are marked SafeToRemove or MigrationOnly - so it only returns fields that are represented in the Haskell type. If you need those fields, use getEntityFieldsDatabase.

This returns all of the FieldDef defined for the EntityDef, including those fields that are marked as MigrationOnly (and therefore only present in the database) or SafeToRemove (and a migration will drop the column if it exists in the database).

For all the fields that are present on the Haskell-type, see getEntityFields.

Retrieve the Haskell name of the given entity.