Bulk JSON to CSV — Convert Multiple Files or Large Datasets

A browser-based JSON to CSV converter handles one file at a time. When you have 50 files, 500 files, or a single file that is too large for a browser tab, you need a different approach. Here are the actual scripts and commands for batch conversion — ready to copy and run.

When to Use a Browser Tool vs Scripts

Scenario	Browser Tool	Script / CLI
1 file, under 50MB	✓ Fastest option	Overkill
1 file, 50-100MB	✓ Works on most devices	~Not needed unless slow
1 file, 100MB+	✗ May crash browser	✓ Use jq or Python
5-10 files	~Tedious but doable	✓ Worth scripting
50+ files	✗ Not practical	✓ Definitely script it
Daily automated conversion	✗ Cannot automate	✓ Cron job or CI/CD
Custom field extraction	✗ Converts everything	✓ Full control with jq/pandas
Sensitive data check	✓ Data stays local	✓ Also local

Method 1: Bash + jq (Fastest for Simple JSON)

Requirements: bash (Mac/Linux/WSL), jq installed (brew install jq or apt install jq).

Convert a single file — flat JSON array:

If your JSON is an array of objects with the same keys:

• jq -r '(.[0] | keys_unsorted) as $keys | $keys, map([.[ $keys[] ]])[] | @csv' input.json > output.csv

This extracts column headers from the first object, then maps every object to CSV rows. Handles any number of fields automatically.

Batch convert all JSON files in a directory:

• for f in *.json; do jq -r '(.[0] | keys_unsorted) as $keys | $keys, map([.[ $keys[] ]])[] | @csv' "$f" > "${f%.json}.csv"; done

This loops through every .json file, converts it, and saves a .csv with the same name. A directory of 100 files converts in seconds.

Extract specific fields only:

• jq -r '.[] | [.name, .email, .address.city] | @csv' input.json > output.csv

This picks only name, email, and city (flattening the address object manually). Add a header row by echoing column names first.

Method 2: Python + pandas (Best for Nested JSON)

Requirements: Python 3.x, pandas (pip install pandas).

Single file conversion:

• Import json and pandas
• Load: data = json.load(open("input.json"))
• Normalize: df = pd.json_normalize(data)
• Save: df.to_csv("output.csv", index=False)

json_normalize automatically flattens nested objects with dot-notation column names. This is the most reliable method for complex nested JSON.

Batch convert a directory:

• Use glob.glob("*.json") to find all JSON files
• Loop through each file, load, normalize, save with .replace(".json", ".csv")
• Optionally add a source_file column to track which file each row came from

Merge all JSON files into one CSV:

• Load all files into a list of DataFrames
• Use pd.concat(frames, ignore_index=True) to merge
• Missing columns across files become NaN (handled automatically)
• Save the combined DataFrame to one CSV

Method 3: Node.js (For JS/TS Projects)

Requirements: Node.js, npm install json2csv

• Import fs, path, and json2csv
• Read JSON with JSON.parse(fs.readFileSync("input.json"))
• Convert with new Parser({ flattenObjects: true }).parse(data)
• Write with fs.writeFileSync("output.csv", csv)

The json2csv library handles flattening, escaping, and header generation. The flattenObjects: true option creates dot-notation columns for nested objects — same behavior as the browser converter and pandas.

Handling Large Files (500MB+)

Large JSON files crash browser tabs and exhaust Node.js memory. Streaming solutions:

• jq: Streams by default. Even multi-GB files process with constant memory. It reads the JSON incrementally and outputs CSV rows as it goes.
• Python ijson: pip install ijson. Parses JSON incrementally: ijson.items(open("big.json", "rb"), "item") yields one object at a time. Write each to CSV as you go.
• JSON Lines (.jsonl): If you control the source format, use JSONL instead of JSON arrays. Each line is an independent object, enabling line-by-line processing with zero memory overhead.

File Size	Recommended Tool	Memory Usage	Notes
Under 50MB	Browser converter	~100-200MB (browser tab)	Fastest for one-off work
50-500MB	jq or Python pandas	~Same as file size	pandas loads into memory; jq streams
500MB-5GB	jq (streaming)	~Constant (few MB)	jq processes incrementally
500MB-5GB	Python ijson	~Constant (few MB)	Streaming parser, process row by row
5GB+	jq or custom streaming	~Constant	Split file first if possible

When the Browser Tool Still Helps

Even when you are doing batch work, the browser converter has a role:

• Validating structure before scripting: Paste a sample from one file to see how the converter flattens it. Use that column structure as the template for your script.
• Spot-checking results: After batch conversion, paste a random file into the browser tool and compare output to your script's output. Quick sanity check.
• Converting outliers: If 98 of 100 files follow the same schema but 2 are different, batch-convert the 98 and manually handle the 2 in the browser.
• Sharing with non-technical teammates: You built the pipeline, but your colleague just has one file. Send them the converter link instead of teaching them jq.

Related Tools

• JSON to CSV Converter — single-file browser conversion
• JSON Formatter — validate JSON structure before batch processing
• Merge CSV — combine multiple output CSV files
• Remove Duplicate Rows — deduplicate after merging
• CSV Sanitizer — clean messy data in converted output
• CSV to JSON — reverse direction when feeding data back to APIs
• CSV Deduplicator — remove duplicate entries across merged files
• Code Formatter — format your conversion scripts cleanly