Working with Files¶

jpx provides several ways to read input and write output, making it flexible for different workflows.

Reading Input¶

From stdin (default)¶

Pipe JSON directly to jpx:

echo '{"name": "alice"}' | jpx 'name'
# "alice"

cat data.json | jpx 'users[*].name'

curl -s https://api.example.com/data | jpx 'items[0]'

From a File¶

Pass the file as a trailing argument (jq-style):

jpx 'users[*].name' data.json

jpx 'settings.theme' config.json

Or use -f / --file explicitly:

jpx -f data.json 'users[*].name'

jpx --file config.json 'settings.theme'

Null Input¶

Use -n or --null-input for expressions that don't need input data:

jpx -n 'now()'
# "2024-01-15T10:30:00Z"

jpx -n 'uuid()'
# "550e8400-e29b-41d4-a716-446655440000"

jpx -n 'range(1, 5)'
# [1, 2, 3, 4]

Writing Output¶

To stdout (default)¶

Output goes to stdout with pretty-printing by default:

jpx -f data.json 'users[0]'
# {
#   "name": "alice",
#   "age": 30
# }

To a File¶

Use -o or --output to write to a file:

jpx -f data.json 'users[?active]' -o active-users.json

jpx -f input.json 'transform(@)' --output result.json

Compact Output¶

Use -c or --compact for single-line JSON (no pretty-printing):

jpx -c -f data.json 'users[0]'
# {"name":"alice","age":30}

Processing Multiple Files¶

Sequential Processing¶

Process files one at a time in a loop:

for f in *.json; do
  jpx -f "$f" 'metadata.title'
done

Combining Files with Slurp¶

Use -s or --slurp to combine multiple JSON values into an array:

cat file1.json file2.json file3.json | jpx -s 'length(@)'
# 3

cat *.json | jpx -s '[*].name'

Streaming JSON Lines (NDJSON)¶

Use --stream (or --each) for newline-delimited JSON:

# Process each line independently
cat events.ndjson | jpx --stream 'event_type'

# Filter and transform streaming data
cat logs.ndjson | jpx --stream '[?level == `error`].message'

Streaming mode uses constant memory regardless of file size.

Query Files¶

Store queries in files for reuse:

Plain Text Query Files¶

# query.txt contains: users[?active].{name: name, email: email}
jpx -Q query.txt -f data.json

Query Libraries (.jpx)¶

Create reusable query collections:

-- queries.jpx
-- :name active-users
-- :description Get all active users
users[?active]

-- :name user-emails
-- :description Extract just the email addresses
users[*].email

Run named queries:

jpx -Q queries.jpx:active-users -f data.json

# Or with --query flag
jpx -Q queries.jpx --query user-emails -f data.json

# List available queries
jpx -Q queries.jpx --list-queries

See Query Files for more details.

JSON Patch Operations¶

Generate Patches¶

Compare two files and generate a JSON Patch (RFC 6902):

jpx --diff original.json modified.json > changes.patch

Apply Patches¶

Apply a patch to a document:

jpx --patch changes.patch -f document.json -o updated.json

Merge Patches¶

Apply a JSON Merge Patch (RFC 7396):

jpx --merge updates.json -f document.json

Tips¶

Checking File Validity¶

Use --stats to quickly validate and analyze a JSON file:

jpx --stats -f data.json

Exploring File Structure¶

Use --paths to see all paths in a JSON document:

jpx --paths -f data.json

# With types
jpx --paths --types -f data.json

# With values
jpx --paths --values -f data.json

Benchmarking File Operations¶

Measure query performance on a file:

jpx --bench 1000 -f large-data.json 'complex[*].transformation'