by Hugo Di Francesco

A gentle introduction to GraphQL API integrations

Annie Spratt on Unsplash

GraphQL is a great alternative to REST (or other HTTP API designs). This is a quick introduction to the core concepts around consuming a GraphQL API.

To see some examples consuming a GraphQL API:

What is GraphQL and what problems does it solve?

GraphQL is “a query language for your API”.

In plain English, it makes the client define what (nested) data it needs.

If we compare it to REST approaches:

  • the “pure” REST approach is to return IDs (or resource links) for any associations (or nested resources).
  • The less pure approach is to expand all the nested stuff.

The first situation leads to having to make lots of calls to fetch all the data. The second leads to huge payloads and slow load times.

In GraphQL, the client states in the request what it wants expanded, renamed or whatever else in the response.

It has some nice side-effects, for example less need to version your API since the client defines what it wants and GraphQL has a way to deprecate fields.


GraphiQL, “An in-browser IDE for exploring GraphQL.” is available by navigating to the endpoint in your browser. It’s possible to generate the schema using the GraphQL CLI (requires Node + npm 5+):

npx graphql-cli get-schema --endpoint $BASE_URL/api/graphql --no-all -o schema.graphql


GraphQL query concepts


What we would like returned in the query, see the GraphQL documentation for “fields”. The GraphQL query for that returns the fields name, fleeRate, maxCP, maxHP, is the following:

{  pokemon(name: "Pikachu") {    name    fleeRate    maxCP    maxHP  }}


How we are going to filter the query data down, see the GraphQL documentation for “arguments”. To get the names of the first 10 pokemon we use pokemons (first: 10) { FIELDS }to see the output here:

{  pokemons (first: 10) {    name    fleeRate    maxCP    maxHP  }}


Aliases give us the ability to rename fields. (See the GraphQL documentation for “aliases”). We’re actually going to use it to map fields in the query eg. from camel to snake case:

{  pokemon(name: "Pikachu") {    evolution_requirements: evolutionRequirements {      amount      name    }  }}

Running this query (here) gives us the following, where the evolutionRequirements is what we’ve aliased it to.

{  "data": {    "pokemon": {      "evolution_requirements": {        "amount": 50,        "name": "Pikachu candies"      }    }  }}


The definition of fields to be expanded on a type. It’s a way to keep the queries DRY and in general split out the field definitions that are repeated, re-used or deeply nested, see the GraphQL documentation for fragments. It’s going to mean that instead of doing (see the query in action here):

{  pokemon(name: "Pikachu") {    weight {      minimum      maximum    }    height {      minimum      maximum    }  }}

We can for example run this (query here):

{  pokemon(name: "Pikachu") {    weight {...FullPokemonDimensions}    height {...FullPokemonDimensions}  }}
fragment FullPokemonDimensions on PokemonDimension {  minimum  maximum}

The output is equivalent:

{  "data": {    "pokemon": {      "weight": {        "minimum": "5.25kg",        "maximum": "6.75kg"      },      "height": {        "minimum": "0.35m",        "maximum": "0.45m"      }    }  }}

Running a GraphQL query

A GraphQL query can be run over POST or GET, it consists of:

POST (recommended)

  • Required headers: Content-Type: application/json
  • Required JSON body parameter: query: { # insert your query }

Raw HTTP request

POST / HTTP/1.1Host: application/json
{        "query": "{ pokemons(first: 10) { name } }"}


curl -X POST \ \  -H 'Content-Type: application/json' \  -d '{        "query": "{ pokemons(first: 10) { name } }"    }'


  • Required query param: query

raw HTTP request

GET /?query={%20pokemons(first:%2010)%20{%20name%20}%20} HTTP/1.1Host:


curl -X GET '{%20pokemons%28first:%2010%29%20{%20name%20}%20}'

Top-level queries

There are 2 types of queries on the GraphQL Pokemon API at the moment:

  • First X pokemon: get all items (with whatever fields are defined in the query)
  • Single Pokemon by name: get a single item by its slug (with whatever fields are defined in the query)
  • Single Pokemon by id: get a single item by its slug (with whatever fields are defined in the query)

First X Pokemon

Queries of the form (see it in action in GraphiQL):

{  pokemons(first: 5) {    name    # other fields  }}

Single Pokemon by name

Queries of the form (see it in action in GraphiQL):

{  pokemon(name: "Pikachu") {    name    classification    # other fields  }}
Note the double quotes ("") around the argument value

Single Pokemon by id

Queries of the form (see it in action in GraphiQL):

{  pokemon(id: "UG9rZW1vbjowMjU=") {    name    classification    # other fields  }}
Note the double quotes ("") around the argument value

Sample queries

Get some Pokemon to create strengths/weakness/resistance classification

Query (see it in GraphiQL):

{  pokemons(first: 100) {    name    image    maxHP    types    weaknesses    resistant  }}

Get Pokemon and evolutions expanded for physical stats and attacks

Query (see it in GraphiQL):

{  pokemon(name: "Pikachu") {    ...PokemonWithAttack    ...FullPhysicalStats    evolutions {      ...FullPhysicalStats      ...PokemonWithAttack    }  }}
fragment PokemonWithAttack on Pokemon {  name  attacks {    fast {      name      type      damage    }    special {      name      type      damage    }  }}
fragment FullPhysicalStats on Pokemon {  height { ...FullDimension }  weight { ...FullDimension }}
fragment FullDimension on PokemonDimension {  minimum  maximum}

Get selected Pokemon as named fields with their evolution names

Query (see it in GraphiQL).

We can rename top-level queries using aliases. That’s helpful if we want to do the following:

{  pikachu: pokemon(name: "Pikachu") {    ...FullPokemon    evolutions {      ...FullPokemon    }  }  bulbasaur:pokemon(name: "Bulbasaur") {    ...FullPokemon    evolutions {      ...FullPokemon    }  }}
fragment FullPokemon on Pokemon {  name}

If you want to learn how to integrate with a GraphQL API:

- In Python, see Python GraphQL client requests example using gql
- In JavaScript browser and Node, see last week’s Code with Hugo newsletter

Originally published at on September 10, 2018.