JSONPath vs JQ: When to Use Which

Both JSONPath and jq let you pull fields from nested JSON without writing imperative loops. JSONPath feels familiar if you know XPath; jq is a full data-processing language with filters, pipes, and transformations. Choosing the wrong one usually means either fighting the syntax or over-engineering a simple extraction.

What JSONPath is good at

JSONPath (implementations vary slightly) uses path expressions to select nodes:

• Root: $
• Child: $.store.book
• Index: $.items[0]
• Wildcard: $.store.*
• Recursive descent: $..author
• Filter (where supported): $.books[?(@.price < 10)]

Best for:

• Quick field lookups in API responses ($.data.user.email)
• Config templates that document paths for non-programmers
• Tools that already speak JSONPath (Postman tests, some API gateways)

Limitations:

• Not a programming language — complex transforms (group-by, map-reduce) get awkward
• Filter syntax differs between implementations
• No built-in string/math operations in the core spec

What jq is good at

jq uses a pipeline model: input | filter | filter. Example:

.users[] | select(.active == true) | { id, name: .profile.displayName }

Best for:

• Shaping API responses for frontends (pick, rename, nest)
• Aggregations: group_by, map, add, unique
• Cleaning messy arrays before validation
• Shell scripts: curl ... | jq '.items[].id'

Limitations:

• Steeper learning curve than JSONPath
• Heavier runtime (BracketView runs jq via WebAssembly in the browser)
• Overkill for $.id style lookups

Side-by-side comparison

| Task | JSONPath | jq | |------|----------|-----| | Get nested field | $.order.customer.id | .order.customer.id | | All items in array | $.items[*] | .items[] | | Filter array | $.items[?(@.qty > 0)] | .items[] \| select(.qty > 0) | | Rename keys in output | Awkward | { userId: .id, ... } | | Count / sum | Not native | map(.qty) \| add |

When to use JSONPath

• Your team or tool chain already documents paths in JSONPath notation.
• You need a one-line extraction with minimal syntax to learn.
• You are writing tests or docs that non-jq users must read.

When to use jq

• You need to transform structure, not just read values.
• You are piping CLI output or building ETL-style steps.
• Filters depend on multiple fields or computed conditions.

Running both in BracketView

• JSONPath Query — paste JSON, run path expressions, inspect matches in the tree.
• JQ Playground — full jq engine in the browser; iterate on pipelines with live output.

Start with JSONPath for simple reads; reach for jq when you catch yourself wanting "map this array and group by status."