8. Statement Fragments

In this chapter we discuss how to construct SQL statements at runtime.

Setting Up

Same as last chapter, so if you’re still set up you can skip this section. Otherwise let’s set up a Transactor and YOLO mode.

import doobie.imports._
import cats._, cats.data._, cats.implicits._
import fs2.interop.cats._
val xa = DriverManagerTransactor[IOLite](
  "org.postgresql.Driver", "jdbc:postgresql:world", "postgres", ""
)
val y = xa.yolo; import y._

We’re still playing with the country table, shown here for reference.

CREATE TABLE country (
  code       character(3)  NOT NULL,
  name       text          NOT NULL,
  population integer       NOT NULL,
  gnp        numeric(10,2)
  -- more columns, but we won't use them here
)

Composing SQL literals

SQL literals constructed with the fr interpolator behave just like the sql interpolator and can be composed by concatenation.

scala> val a = fr"select name from country"
a: doobie.util.fragment.Fragment = Fragment("select name from country ")

scala> val b = fr"where code = 'USA'"
b: doobie.util.fragment.Fragment = Fragment("where code = 'USA' ")

scala> val c = a ++ b // concatenation by ++
c: doobie.util.fragment.Fragment = Fragment("select name from country where code = 'USA' ")

scala> c.query[String].unique.quick.unsafePerformIO
  United States

Fragments can capture arguments of any type with a Param instance, just as the sql interpolator does.

scala> def whereCode(s: String) = fr"where code = $s"
whereCode: (s: String)doobie.util.fragment.Fragment

scala> val fra = whereCode("FRA")
fra: doobie.util.fragment.Fragment = Fragment("where code = ? ")

scala> (fr"select name from country" ++ fra).query[String].quick.unsafePerformIO
  France

You can lift an arbitrary string value via Fragment.const, which allows you to parameterize on things that aren’t valid SQL parameters.

scala> def count(table: String) = (fr"select count(*) from" ++ Fragment.const(table)).query[Int].unique
count: (table: String)doobie.free.connection.ConnectionIO[Int]

scala> count("city").quick.unsafePerformIO
  4079

Whitespace handling

The rendered SQL string for a fr or const fragment will have a single space character appended, which is usually what you want. Normally you don’t need to worry about whitespace when composing fragments.

If you do not want a fragment to have trailing space you can use the fr0 interpolator or const0 constructor. This is used here and there in the Fragments module to yield prettier SQL strings.

scala> fr"IN (" ++ List(1, 2, 3).map(n => fr"$n").intercalate(fr",") ++ fr")"
res3: doobie.util.fragment.Fragment = Fragment("IN ( ? , ? , ? ) ")

scala> fr0"IN (" ++ List(1, 2, 3).map(n => fr0"$n").intercalate(fr",") ++ fr")"
res4: doobie.util.fragment.Fragment = Fragment("IN (?, ?, ?) ")

Note that the sql interpolator is simply an alias for fr0.

The Fragments Module

The Fragments module provides some combinators for common patterns when working with fragments. The following example illustrates a few of them. See the Scaladoc or source for more information.

Here we define a query with a three optional filter conditions.

// Import some convenience combinators.
import Fragments.{ in, whereAndOpt }

// Country Info
case class Info(name: String, code: String, population: Int)

// Construct a Query0 with some optional filter conditions and a configurable LIMIT.
def select(name: Option[String], pop: Option[Int], codes: List[String], limit: Long) = {

  // Three Option[Fragment] filter conditions.
  val f1 = name.map(s => fr"name LIKE $s")
  val f2 = pop.map(n => fr"population > $n")
  val f3 = codes.toNel.map(cs => in(fr"code", cs))

  // Our final query
  val q: Fragment =
    fr"SELECT name, code, population FROM country" ++
    whereAndOpt(f1, f2, f3)                         ++
    fr"LIMIT $limit"

  // Consruct a Query0
  q.query[Info]

}

We first construct three optional filters, the third of which uses the in combinator to construct an SQL IN clause. The final statement uses the whereAndOpt combinator that constructs a WHERE clause with the passed sequence of Option[Fragment] joined with AND if any are defined, otherwise it evaluates to the empty fragment. The end result is that the WHERE clause appears only if at least one filter is defined.

Let’s look at a few possibilities.

scala> select(None, None, Nil, 10).check.unsafePerformIO // no filters

  SELECT name, code, population FROM country LIMIT ? 

   SQL Compiles and Typechecks
   P01 Long    BIGINT (int8)
   C01 name       VARCHAR (varchar) NOT NULL    String
   C02 code       CHAR    (bpchar)  NOT NULL    String
   C03 population INTEGER (int4)    NOT NULL    Int

scala> select(Some("U%"), None, Nil, 10).check.unsafePerformIO // one filter

  SELECT name, code, population FROM country WHERE name LIKE ? LIMIT ? 

   SQL Compiles and Typechecks
   P01 String    VARCHAR (text)
   P02 Long      BIGINT  (int8)
   C01 name       VARCHAR (varchar) NOT NULL    String
   C02 code       CHAR    (bpchar)  NOT NULL    String
   C03 population INTEGER (int4)    NOT NULL    Int

scala> select(Some("U%"), Some(12345), List("FRA", "GBR"), 10).check.unsafePerformIO // three filters

  SELECT name, code, population FROM country WHERE name LIKE ? AND population > ? AND code IN (?, ?) LIMIT ? 

   SQL Compiles and Typechecks
   P01 String    VARCHAR (text)
   P02 Int       INTEGER (int4)
   P03 String    CHAR    (bpchar)
   P04 String    CHAR    (bpchar)
   P05 Long      BIGINT  (int8)
   C01 name       VARCHAR (varchar) NOT NULL    String
   C02 code       CHAR    (bpchar)  NOT NULL    String
   C03 population INTEGER (int4)    NOT NULL    Int