duckdb-simple

duckdb-simple provides a high-level Haskell interface to DuckDB inspired by the APIs of sqlite-simple and postgresql-simple. It builds on the low-level bindings exposed by duckdb-ffi and provides a focused API for opening connections, running queries, binding parameters, and decoding typed results—including the full set of DuckDB scalar types (signed/unsigned integers, decimals, hugeints, intervals, precise and timezone-aware temporals, blobs, enums, bit strings, and bignums).

Getting Started

{-# LANGUAGE OverloadedStrings #-}

import Database.DuckDB.Simple
import Database.DuckDB.Simple.Types (Only (..))

main :: IO ()
main =
  withConnection ":memory:" \conn -> do
    _ <- execute_ conn "CREATE TABLE items (id INTEGER, name TEXT)"
    _ <- execute conn "INSERT INTO items VALUES (?, ?)" (1 :: Int, "banana" :: String)
    rows <- query_ conn "SELECT id, name FROM items ORDER BY id"
    mapM_ print (rows :: [(Int, String)])

Key Modules

• Database.DuckDB.Simple – connections, prepared statements, execution, queries, metadata, and error handling.
• Database.DuckDB.Simple.ToField / ToRow – typeclasses and helpers for preparing positional or named parameters.
• Database.DuckDB.Simple.FromField / FromRow – typeclasses for decoding query results, with generic deriving support for product types.
• Database.DuckDB.Simple.Generic – automatic encoding/decoding of Haskell ADTs as DuckDB STRUCTs and UNIONs via GHC generics and the ViaDuckDB deriving-via helper.
• Database.DuckDB.Simple.LogicalRep – structured value types (StructValue, UnionValue) for working with DuckDB's composite types.
• Database.DuckDB.Simple.Types – shared types (Query, Null, Only, (:.), SQLError).
• Database.DuckDB.Simple.Function – register scalar Haskell functions that can be invoked directly from SQL.

Querying Data

import Database.DuckDB.Simple
import Database.DuckDB.Simple.Types (Only (..))

fetchNames :: Connection -> IO [Maybe String]
fetchNames conn = do
  _ <- execute_ conn "CREATE TABLE names (value TEXT)"
  _ <- executeMany conn "INSERT INTO names VALUES (?)"
    [Only (Just "Alice"), Only (Nothing :: Maybe String)]
  fmap fromOnly <$> query_ conn "SELECT value FROM names ORDER BY value IS NULL, value"

The execution helpers return the number of affected rows (Int) so callers can assert on data changes when needed.

Named Parameters

duckdb-simple supports both positional (?) and named parameters. Named parameters are bound with the (:=) helper exported from Database.DuckDB.Simple.ToField.

import Database.DuckDB.Simple
import Database.DuckDB.Simple.ToField (NamedParam ((:=)))

insertNamed :: Connection -> IO Int
insertNamed conn =
  executeNamed conn
    "INSERT INTO events VALUES ($kind, $payload)"
    ["$kind" := ("metric" :: String), "$payload" := ("ok" :: String)]

DuckDB does not allow mixing positional and named placeholders within the same SQL statement; the library preserves DuckDB’s error message in that situation. DuckDB does not support savepoints. This library does not provide withSavepoint.

If the number of supplied parameters does not match the statement’s declared placeholders—or if you attempt to bind named arguments to a positional-only statement—duckdb-simple raises a FormatError before executing the query.

Decoding rows

FromRow is powered by a RowParser, which means instances can be written in a monadic/Applicative style and even derived generically for product types:

{-# LANGUAGE DeriveAnyClass #-}
{-# LANGUAGE DeriveGeneric #-}

import Database.DuckDB.Simple
import GHC.Generics (Generic)

data Person = Person
  { personId :: Int
  , personName :: Text
  }
  deriving stock (Show, Generic)
  deriving anyclass (FromRow)

fetchPeople :: Connection -> IO [Person]
fetchPeople conn = query_ conn "SELECT id, name FROM person ORDER BY id"

Helper combinators such as field, fieldWith, and numFieldsRemaining are available when a custom instance needs fine-grained control.

Generic Encoding with ViaDuckDB

The Database.DuckDB.Simple.Generic module provides automatic encoding and decoding of Haskell algebraic data types as DuckDB STRUCTs and UNIONs via GHC generics.

Product Types as STRUCTs

Product types (records) are automatically encoded as DuckDB STRUCT values:

{-# LANGUAGE DeriveGeneric #-}
{-# LANGUAGE DerivingVia #-}

import Data.Int (Int64)
import Data.Text (Text)
import Database.DuckDB.Simple
import Database.DuckDB.Simple.Generic (ViaDuckDB (..))
import GHC.Generics (Generic)

data User = User
  { userId :: Int64
  , userName :: Text
  }
  deriving stock (Eq, Show, Generic)
  deriving (DuckDBColumnType, ToField, FromField) via (ViaDuckDB User)

-- Round-trip through the database
storeAndFetchUser :: Connection -> User -> IO [User]
storeAndFetchUser conn user = do
  _ <- execute_ conn "CREATE TABLE users (data STRUCT(userId BIGINT, userName TEXT))"
  _ <- execute conn "INSERT INTO users VALUES (?)" (Only user)
  fmap fromOnly <$> query_ conn "SELECT data FROM users"

Sum Types as UNIONs

Sum types are encoded as DuckDB UNION values, with each constructor becoming a union member:

data Shape
  = Circle Double
  | Rectangle Double Double
  | Point
  deriving stock (Eq, Show, Generic)
  deriving (DuckDBColumnType, ToField, FromField) via (ViaDuckDB Shape)

-- Store and retrieve shape data
storeShape :: Connection -> Shape -> IO [Shape]
storeShape conn shape = do
  _ <- execute_ conn
    "CREATE TABLE shapes (s UNION(Circle STRUCT(field1 DOUBLE), \
    \Rectangle STRUCT(field1 DOUBLE, field2 DOUBLE), Point STRUCT()))"
  _ <- execute conn "INSERT INTO shapes VALUES (?)" (Only shape)
  fmap fromOnly <$> query_ conn "SELECT s FROM shapes"

Nullary constructors (like Point) are encoded with a null payload. Non-record constructors use positional field names (field1, field2, etc.).

Arrays and Lists

DuckDB arrays (fixed-length) and lists (variable-length) are also supported:

import Data.Array (Array, listArray)

storeArray :: Connection -> IO [Array Int Int]
storeArray conn = do
  _ <- execute_ conn "CREATE TABLE arrays (vals INTEGER[3])"
  let arr = listArray (0, 2) [1, 2, 3] :: Array Int Int
  _ <- execute conn "INSERT INTO arrays VALUES (?)" (Only arr)
  fmap fromOnly <$> query_ conn "SELECT vals FROM arrays"

storeList :: Connection -> IO [[Int]]
storeList conn = do
  _ <- execute_ conn "CREATE TABLE lists (vals INTEGER[])"
  let arr = listArray (0, 2) [1, 2, 3] :: Array Int Int
  _ <- execute conn "INSERT INTO lists VALUES (?)" (Only arr)
  fmap fromOnly <$> query_ conn "SELECT vals FROM lists"

A Haskell list reads a LIST result. Lists have no parameter instance, so bind an Array. DuckDB casts the array to the LIST column type.

An array parameter with scalar elements takes its element type from the column type name of the element, so an empty array keeps its type. An array of STRUCT values, UNION values, generic records, or arrays takes the type of its first element that is not NULL. All other non-NULL elements must have the same type, including field names, decimal precision and scale, and nested types. Different types raise an error before DuckDB can cast away fields or round values. An empty or all-NULL array of such elements raises an error.

Infinite dates and timestamps

Use Database.DuckDB.Simple.Time when a column can contain temporal infinity. Its Date, LocalTimestamp, and UTCTimestamp types wrap Day, LocalTime, and UTCTime in Unbounded: NegInfinity, Finite value, or PosInfinity. These types support parameters, results, and fields in generic composites. UTCTimestamp binds as TIMESTAMPTZ.

import Database.DuckDB.Simple.Time

infiniteDates :: Connection -> IO [Only Date]
infiniteDates conn = query conn "SELECT ?::DATE" (Only (PosInfinity :: Date))

The ordinary Day, LocalTime, and UTCTime instances reject infinity with a conversion error. Use Maybe Date to distinguish SQL NULL from infinity. Floating-point NaN and infinities remain valid Float and Double values.

Manual STRUCT and UNION Handling

Temporal fields retain their SQL units when composite values are rebound. The FieldDate, FieldTimestamp, and FieldTimestampTZ constructors hold Unbounded values. Wrap finite payloads in Finite when constructing them. For TIMESTAMP_S or TIMESTAMP_MS values outside the TIMESTAMP range, use an explicit parameter cast, such as SELECT ?::STRUCT(value TIMESTAMP_S). DuckDB otherwise attempts to convert these parameters to microseconds.

For more control, you can work directly with StructValue and UnionValue from Database.DuckDB.Simple.LogicalRep:

import Database.DuckDB.Simple.LogicalRep (StructValue (..), UnionValue (..))
import Database.DuckDB.Simple.FromField (FieldValue (..))

manualStruct :: Connection -> IO [(StructValue FieldValue, UnionValue FieldValue)]
manualStruct conn = do
  _ <- execute_ conn
    "CREATE TABLE composite (s STRUCT(a INT, b INT), \
    \u UNION(x INT, y VARCHAR))"
  [(s, u)] <- query_ conn
    "SELECT {'a': 1, 'b': 2}, \
    \CAST(union_value(x := 42) AS UNION(x INT, y VARCHAR))"
    :: IO [(StructValue FieldValue, UnionValue FieldValue)]
  _ <- execute conn "INSERT INTO composite VALUES (?, ?)" (s, u)
  query_ conn "SELECT s, u FROM composite"

Resource Management

• withConnection and withStatement wrap the open/close lifecycle and guard against exceptions; use them whenever possible to avoid leaking C handles.
• All intermediate DuckDB objects (results, prepared statements, values) are released immediately after use. Query helpers return a Haskell list of all rows. Folds and cursors decode rows incrementally, while DuckDB retains the materialized native result until it is exhausted, reset, or closed.
• execute/query variants reset statement bindings each run so prepared statements can be reused safely.

For concurrent workers, use a separate connection per worker. A connection can move between threads or be shared when the application serializes access. Hold that lock for the whole transaction or cursor lifetime, including close. DuckDB serializes native query calls, but this does not protect the Haskell handle state or prevent another call from interfering with an active cursor. Statements and connections do not provide their own lock. A callback must not execute another query on its active connection or close that connection.

Link your executable with ghc-options: -threaded to allow prompt cancellation of native queries, including Ctrl-C. On cancellation, the library interrupts DuckDB and waits for the native call to return before it releases resources and propagates the exception. Cancellation is cooperative: native code and Haskell callbacks must return before cleanup can finish.

Close or reset an abandoned statement to release its result. DuckDB can retain native result buffers until that result is destroyed. Use withStatement for manual iteration. fold releases the result after success or an exception. The accumulator determines Haskell memory use.

Metadata helpers

• columnCount and columnName expose prepared-statement metadata so you can inspect result shapes before executing a query.
• execute returns the number of affected rows. Use SQL RETURNING clauses when you need generated identifiers.

Cursors and folds

fold, fold_, and foldNamed decode one row at a time from DuckDB's result chunks. DuckDB 1.5 materializes the native result before the first row is returned. These functions avoid a complete Haskell row list, but native memory use still depends on the result size. The native API for starting a streaming result is deprecated; the default interface uses the supported execution API.

import Database.DuckDB.Simple.Types (Only (..))

sumValues :: Connection -> IO Int
sumValues conn =
  fold_ conn "SELECT n FROM stream_fold ORDER BY n" 0 $ \acc (Only n) ->
    pure (acc + n)

For manual cursor-style iteration, use nextRow/nextRowWith on an open Statement to pull rows one at a time and decide when to stop.

Cursors support the same column types as eager queries, including STRUCT and UNION values with nested collections and NULLs. VARIANT and GEOMETRY work with eager queries, cursors, and folds.

Optional native streaming

Database.DuckDB.Simple.Deprecated.Streaming provides fold, fold_, foldNamed, nextRow, and nextRowWith with native streaming enabled. Import it qualified:

import qualified Database.DuckDB.Simple.Deprecated.Streaming as Streaming

streamSum :: Connection -> IO Int
streamSum conn =
  Streaming.fold_ conn "SELECT i FROM range(1000000) t(i)" 0 $ \acc (Only n) ->
    pure (acc + n)

The import emits a deprecation warning because DuckDB has deprecated the execution entry point. DuckDB can still materialize some queries. Streaming does not bound the memory used by query operators.

The first cursor fetch selects the execution mode until an explicit reset. Switching between default and streaming nextRow calls retains that mode. Keep the connection dedicated to the active stream; another query on that connection can invalidate it. Cancellation interrupts native chunk fetching as well as execution.

This module also provides foldArrow and foldArrow_ for streaming Arrow batches. They use the same supported Arrow conversion and scoped ownership as Database.DuckDB.Simple.Arrow.

Arrow batches

Database.DuckDB.Simple.Arrow provides foldArrow and foldArrow_ for clients that consume the Arrow C Data Interface. Each callback receives a separate schema and array batch. It can read them or pass them to an Arrow consumer that releases or moves them. The fold releases any remaining contents on success, failure, or cancellation. The consumer owns any contents it moves.

The original pointers are valid only during the callback. To retain contents, a consumer must move the root structs into its own storage and set the source release fields to NULL. Moved contents remain valid after the query and connection close. Empty results do not produce a callback.

For example, the dataframe-arrow-bridge package can copy each batch into a Haskell DataFrame and release the Arrow objects:

import qualified DataFrame.IO.Arrow as DataFrame
import qualified Database.DuckDB.Simple.Arrow as Arrow
import Foreign.Ptr (castPtr)

frames <- Arrow.foldArrow_ conn "SELECT id::BIGINT, name::VARCHAR FROM people" [] $ \acc schema array -> do
  frame <- DataFrame.arrowToDataframe (castPtr schema) (castPtr array)
  pure (frame : acc)
-- Reverse frames to recover the query's batch order.

The bridge currently imports signed 32-bit and 64-bit integers, Float, Double, and text columns. It is a test dependency of this repository; applications that use it must declare their own dependency on dataframe-arrow-bridge.

Arrow export uses DuckDB's schema and chunk conversion API. DuckDB materializes the native result before callbacks start, so its memory use depends on the result size. The older Arrow query and scan bindings remain available through Database.DuckDB.FFI.Deprecated and emit deprecation warnings.

VARIANT

A VARIANT result decodes to the FieldValue of the stored value, with its native type. The usual FromField instances read VARIANT columns, for example as Int64, Text, [a], or a generic record. Arrays decode to FieldList. Objects decode to FieldStruct values whose fields have the VARIANT type, in entry order. variantObject builds such a payload from a list of entries. SQL NULL decodes to FieldNull.

Variant from Database.DuckDB.Simple.Variant wraps a FieldValue. Its FromField instance reads any column. Its ToField instance binds the payload as a VARIANT, so SELECT ? returns a VARIANT:

query conn "SELECT ?" (Only (Variant (FieldList [FieldInt8 1, FieldText "two"])))

A scalar payload keeps its native type, such as TINYINT or DECIMAL(4,2). Time and timestamp payloads bind as microsecond types, or as nanosecond types when they have sub-microsecond digits. So a TIMESTAMP_S result binds back as a TIMESTAMP when its value fits. Wider timestamp values use milliseconds or seconds without losing digits. Lists, arrays, and STRUCT fields bind as VARIANT values. MAP and ENUM payloads raise an error. Object parameters reject duplicate keys, empty keys, and keys that contain NUL. You can also bind a plain value and cast it in SQL, as in ?::VARIANT.

The C API cannot create a usable VARIANT type (#27). Each connection reads the VARIANT type the first time a parameter needs it, and its parameters use that type. The connection reads the type with a query on a separate connection. So the query does not run in your transaction, and it does not end a streaming result. Variant has no ToDuckValue instance, and logicalTypeFromRep raises an error for VARIANT, because neither has a connection.

A GEOMETRY payload decodes to FieldGeometry with raw WKB and no CRS. Import these bytes with ST_GeomFromWKB(?)::VARIANT. A Variant parameter that contains FieldGeometry raises an error, also inside arrays and objects. A TIMETZ payload with an offset in seconds raises an error, as a TIMETZ column does. Results can contain empty or NUL object keys. Text and blob payloads can contain NUL.

DuckDB 1.5 has no C API for reading VARIANT values. The decoder checks the native version and physical schema before it reads the internal representation. It checks payload bounds and rejects unknown tags. This format dependency is limited to the supported DuckDB 1.5 line.

Persistent VARIANT columns require storage format v1.5.0 or later. For a new database, pass [("storage_compatibility_version", "v1.5.0")] to openWithConfig or withConnectionWithConfig. The library does not change an existing database's storage compatibility setting.

GEOMETRY

Use Geometry from Data.Geometry for decoded shapes. The geometry-simple package supplies the type and unboxed coordinate vectors. duckdb-simple supplies its parameter and result instances:

import qualified Data.Geometry as G
import qualified Data.Vector.Unboxed as U
import Database.DuckDB.Simple

let line = G.LineString (G.CoordinatesXY (U.fromList [G.XY 0 0, G.XY 1 2, G.XY 3 4]))
rows <- query conn "SELECT ?" (Only line) :: IO [Only G.Geometry]

Points and coordinate sequences store their XY, XYZ, XYM, or XYZM layout. G.EmptyPoint G.DimXYZ is an empty XYZ point. Use Maybe G.Geometry for SQL NULL. Use decodeWKT from Data.Geometry.WKT to parse text without a database.

G.Geometry has no CRS metadata. Reading it returns the coordinates without their CRS label. Binding it creates a GEOMETRY with no CRS. To attach a label, use ST_SetCRS(?, ?) with a shape and CRS text. This does not transform coordinates.

Use RawGeometry from Database.DuckDB.Simple.Geometry to read WKB and CRS without decoding coordinates. Its rawGeometryWKB and rawGeometryCRS fields own their data. They remain usable after the connection closes. Existing ByteString result decoding also returns WKB.

RawGeometry has no ToField instance. Import its bytes explicitly:

import Database.DuckDB.Simple.Geometry

[Only raw] <- query_ conn "SELECT 'POINT Z (1 2 3)'::GEOMETRY('OGC:CRS84')"
rows <- (case rawGeometryCRS raw of
    Nothing -> query conn
        "SELECT system.main.ST_GeomFromWKB(?)"
        (Only (rawGeometryWKB raw))
    Just crs -> query conn
        "SELECT system.main.ST_SetCRS(system.main.ST_GeomFromWKB(?), ?)"
        (rawGeometryWKB raw, crs)
    ) :: IO [Only RawGeometry]

Use Nothing for no CRS and omit ST_SetCRS in that case. Passing SQL NULL to ST_SetCRS returns a NULL geometry. CRS text can contain OGC:CRS84, a custom name, or a full WKT2/PROJJSON definition. DuckDB can reduce a known CRS definition to its registered identifier. It can also normalize WKB byte order. The qualified function names select DuckDB's built-ins even if a user macro has the same name.

fromRawGeometry decodes the shape and drops CRS metadata. toRawGeometry encodes a shape with no CRS. These helpers follow the geometry-simple contracts. NaN in both WKB point X and Y denotes an empty point. Empty multi-geometries and collections have no stored layout tag in the decoded representation. Keep the raw bytes when these details must survive.

Structured parameters use encodeWKT and DuckDB's native cast. DuckDB 1.5 has no WKB value constructor in its C API. The WKT writer combines layouts in multi-geometries and polygon rings. It fills absent Z or M with NaN. DuckDB's WKT parser limits nesting to 16 levels and rejects empty polygon rings and mixed collection layouts. Explicit WKB import preserves mixed member layouts and native NaN payload bits that WKT cannot retain.

FieldGeometry contains a raw result. Generic STRUCT and UNION parameters that contain non-NULL raw geometry raise an error. Construct those values in SQL with explicit WKB import. Their result decoding retains WKB and CRS, including inside LIST, ARRAY, MAP, STRUCT, and UNION values.

LogicalTypeGeometry describes CRS metadata. The C API cannot create a GEOMETRY type with a CRS. So each connection reads a GEOMETRY type for each CRS in a list, together with the VARIANT type, the first time a parameter needs one of them. The default list holds OGC:CRS84. Composite parameters use these types, so typed NULLs, empty collections, and inactive UNION members keep their CRS. Set the list in ConnectionOptions:

let options = defaultConnectionOptions{connectionGeometryCRS = ["OGC:CRS84", "EPSG:3857"]}
withConnectionWithOptions "shapes.duckdb" options \conn -> ...

A list entry can be an identifier, a custom name, or a full WKT2 or PROJJSON definition. A composite parameter with a CRS that is not in the list binds its geometry members without a CRS. logicalTypeFromRep has no connection, so it creates GEOMETRY without a CRS. To apply a CRS in these cases, insert the value into a column with that CRS, or cast it in SQL:

query conn "SELECT ?::UNION(number BIGINT, shape GEOMETRY('OGC:CRS84'))" (Only value)

The CRS in a cast must be a constant. DuckDB rejects a parameter as a type modifier, so ?::GEOMETRY(?) is not valid. To use a CRS that is known only at run time, write it into the query text as a SQL string literal, and double each single quote. A cast accepts a CRS that DuckDB recognizes, such as OGC:CRS84, or a full WKT2 or PROJJSON definition. Unless an extension recognizes it, DuckDB rejects other identifiers, such as EPSG:4326, and custom names. ST_SetCRS also accepts custom names, but it returns GEOMETRY with no CRS for a NULL input.

A bound composite without a CRS also changes the type of expressions that combine it with CRS data. UNION ALL and COALESCE of GEOMETRY and GEOMETRY('OGC:CRS84') give GEOMETRY with no CRS for all rows. Cast the parameter before you combine it with other values.

See geometry-simple for the seven supported families, construction checks, and codec normalization rules.

Feature Coverage

• Connections, prepared statements, positional/named parameter binding.
• High-level execution (execute*) and eager queries (query*, queryNamed).
• Cursor and fold helpers (fold, foldNamed, fold_, nextRow) that decode native result chunks one row at a time.
• Comprehensive scalar type support: signed/unsigned integers, HUGEINT/UHUGEINT, decimals (with width/scale), intervals, precise and timezone-aware temporals, enums, bit strings, blobs, bignums, and UUIDs.
• Composite types: STRUCTs, UNIONs, LISTs, fixed-length ARRAYs, and MAPs with typed parameters and results.
• Generic encoding/decoding: automatic STRUCT/UNION mapping for Haskell ADTs via GHC generics and the ViaDuckDB deriving-via helper.
• Row decoding via FromField/FromRow, with generic deriving for product types.
• User-defined scalar functions backed by Haskell functions (including IO and nullable arguments).
• Transaction helpers (withTransaction) and metadata accessors (columnCount, columnName).

User-Defined Functions

Scalar Haskell functions can be registered with DuckDB connections and used in SQL expressions. Argument and result types reuse the existing FromField and FunctionResult machinery, so Maybe values and IO actions work out of the box.

import Data.Int (Int64)
import Database.DuckDB.Simple
import Database.DuckDB.Simple.Function (createFunction, deleteFunction)
import Database.DuckDB.Simple.Types (Only (..))

registerAndUse :: Connection -> IO [Only Int64]
registerAndUse conn = do
  createFunction conn "hs_times_two" (\(x :: Int64) -> x * 2)
  result <- query_ conn "SELECT hs_times_two(21)" :: IO [Only Int64]
  deleteFunction conn "hs_times_two"
  pure result

Exceptions raised while the function executes are propagated back to DuckDB as SQLError values, and deleteFunction issues a DROP FUNCTION IF EXISTS statement to remove the registration. DuckDB registers C API scalar functions as internal entries; attempting to drop them this way will yield an error, which the library surfaces as an SQLError.

Tests

The test suite is built with tasty and covers connection management, statement lifecycle, parameter binding, and query execution.

cabal test duckdb-simple-test --test-show-details=direct