Field-level errors in GraphQL - Step-by-Step Execution

Concept Flow - Field-level errors

Client sends GraphQL query

↓

Server processes query

↓

Resolve each field

↓

Field OK

↓

Include data

↓

Build response with data and errors

↓

Send response to client

The server processes each field in the query separately, collecting data or errors per field, then sends a combined response with both data and field-level errors.

Execution Sample

GraphQL

query {
  user(id: "1") {
    name
    email
    posts {
      title
    }
  }
}

A query requesting user name, email, and post titles; some fields may return errors individually.

Execution Table

Step	Field	Action	Result	Error Included
1	user	Resolve user by id	User object found	No
2	name	Resolve name field	"Alice"	No
3	email	Resolve email field	null	Yes: Email not available
4	posts	Resolve posts field	List of posts	No
5	posts.title	Resolve title for each post	["Post 1", "Post 2"]	No
6	Build response	Combine data and errors	Data with null email and error message	Yes
7	Send response	Return to client	Response with data and field-level error	Yes

💡 All fields processed; errors included only for fields that failed

Variable Tracker

Variable	Start	After Step 1	After Step 2	After Step 3	After Step 4	After Step 5	Final
data	{}	{"user": {}}	{"user": {"name": "Alice"}}	{"user": {"name": "Alice", "email": null}}	{"user": {"name": "Alice", "email": null, "posts": []}}	{"user": {"name": "Alice", "email": null, "posts": [{"title": "Post 1"}, {"title": "Post 2"}]}}	{"user": {"name": "Alice", "email": null, "posts": [{"title": "Post 1"}, {"title": "Post 2"}]}}
errors	[]	[]	[]	[{"field": "email", "message": "Email not available"}]	[{"field": "email", "message": "Email not available"}]	[{"field": "email", "message": "Email not available"}]	[{"field": "email", "message": "Email not available"}]

Key Moments - 2 Insights

Why does the 'email' field have a null value but also an error message?

Because the server returns partial data for the user, setting 'email' to null when it cannot resolve it, and includes an error message specifically for that field as shown in step 3 and 6 of the execution_table.

Does an error in one field stop the entire query from returning data?

No, the server resolves each field independently, so errors in one field (like 'email') do not prevent other fields (like 'name' and 'posts') from returning data, as seen in steps 2, 3, and 4.

Visual Quiz - 3 Questions

Look at the execution_table, what is the value of 'email' after step 3?

A "[e-mail]"

B null

C "unknown"

D undefined

💡 Hint

Check the 'Result' column for step 3 in the execution_table.

At which step does the server include an error for the 'email' field?

A Step 3

B Step 2

C Step 5

D Step 7

💡 Hint

Look at the 'Error Included' column in the execution_table for when 'email' has an error.

If the 'posts' field failed to resolve, what would change in the execution_table?

A Only the 'name' field would have an error

B The entire query would fail and return no data

C The 'posts' field would have an error and data would be null for posts

D No errors would be included

💡 Hint

Refer to how errors are handled per field in the execution_table and concept_flow.

Concept Snapshot

Field-level errors in GraphQL mean each field is resolved separately.
If a field fails, its value is null and an error message is included.
Other fields still return data normally.
The response contains both 'data' and 'errors' sections.
This allows partial success with detailed error info.

Full Transcript

In GraphQL, when a client sends a query, the server processes each requested field one by one. If a field resolves successfully, its data is included in the response. If a field fails, the server sets that field's value to null and adds an error message about that field. This way, the client receives as much data as possible along with clear information about which fields had problems. For example, if the 'email' field cannot be retrieved, it will be null in the data, and an error will explain why. Other fields like 'name' and 'posts' still return their data. The final response combines all this information, helping clients handle partial failures gracefully.