xml-to-csv

A lightweight, dependency-free XML to CSV converter written in pure Bash and AWK. No Python, no Node.js, no external libraries — just standard Unix tools available on any Linux or macOS system.

---

Features

• Zero dependencies — runs on any system with bash and awk
• Handles single-line, multi-line, self-closing, and empty XML tags
• Properly escapes double quotes in CSV output (" → "")
• Strips Windows-style line endings (\r) automatically
• Outputs to file or stdout for easy piping
• Clear error messages on stderr, clean data on stdout

---

Requirements

Tool	Notes
bash	Version 3.2+
awk	Any POSIX-compatible awk (gawk, mawk, nawk)
tr	Standard Unix tr utility

No installation needed beyond what ships with your OS.

---

Usage

# Convert to stdout
./xml_to_csv.sh input.xml

# Convert to a file
./xml_to_csv.sh input.xml output.csv

# Pipe into another command
./xml_to_csv.sh input.xml | sort
./xml_to_csv.sh input.xml | grep "John"

---

Example

Given employees.xml:

<?xml version="1.0" encoding="UTF-8"?>
<employees>
  <employee>
    <id>1</id>
    <name>Alice Smith</name>
    <department>Engineering</department>
    <email>alice@example.com</email>
  </employee>
  <employee>
    <id>2</id>
    <name>Bob Jones</name>
    <department>Marketing</department>
    <email>bob@example.com</email>
  </employee>
</employees>

Running:

./xml_to_csv.sh employees.xml

Produces:

id,name,department,email
"1","Alice Smith","Engineering","alice@example.com"
"2","Bob Jones","Marketing","bob@example.com"

---

XML Structure Assumptions

The script expects a simple, flat XML structure:

<root>              ← any name, treated as container
  <record>          ← any name, one per row in CSV
    <field>value</field>
    ...
  </record>
  ...
</root>

• Nested XML is not supported — only the immediate children of each record element are extracted as fields
• Field names are taken from the first record; subsequent records follow the same column order
• Fields missing in later records will output as empty strings

---

Supported Tag Formats

Format	Example	Handled
Single-line	<name>Alice</name>	✅
Multi-line	<bio>line1\nline2</bio>	✅
Self-closing	<nickname/>	✅ (empty value)
Empty tag	<nickname></nickname>	✅ (empty value)
XML declaration	<?xml version="1.0"?>	✅ (skipped)
Comments	<!-- comment -->	✅ (skipped)

---

Error Handling

Errors are written to stderr so they never pollute CSV output when piping:

# Safely redirect data and errors separately
./xml_to_csv.sh input.xml > output.csv 2> errors.log

# Suppress error messages entirely
./xml_to_csv.sh input.xml 2>/dev/null

Error	Cause
Input file not found	The specified XML file does not exist
No records processed	XML was empty, malformed, or had no matching record tags

Exit code is 0 on success, 1 on any error.

---

Limitations

• Not a full XML parser — does not handle XML attributes, namespaces, CDATA sections, or deeply nested structures
• Multi-line field values are collapsed into a single space-separated line
• Field order in the output is determined by the order they appear in the first record

---

License

MIT License. See LICENSE for details.