Splitting Text Files by Lines on Linux: A split Command Guide

The split command is the standard tool for breaking large text files into smaller chunks on Linux. Here’s the basic syntax:

split -l 1024 content.txt splitted-content.txt-

This creates output files named splitted-content.txt-aa, splitted-content.txt-ab, and so on, with each containing 1024 lines.

Understanding output naming

By default, split appends two-character suffixes starting with aa. The naming scheme continues alphabetically:

• splitted-content.txt-aa (lines 1–1024)
• splitted-content.txt-ab (lines 1025–2048)
• splitted-content.txt-ac (lines 2049–3072)

If you need more than 676 splits (26² combinations), increase the suffix length with -a:

split -l 1024 -a 3 content.txt splitted-content.txt-

This generates three-character suffixes: aaa, aab, aac, etc.

Numeric suffixes for easier processing

For cleaner sorting and scripting, use numeric suffixes with -d:

split -l 1024 -d content.txt splitted-content.txt-

Output files become: splitted-content.txt-00, splitted-content.txt-01, splitted-content.txt-02, etc.

This is particularly useful when processing files programmatically, since numeric sorting is more predictable across different locales and shells.

Splitting by file size instead of lines

When line count is irrelevant or you’re working with mixed content, use -b to split by byte size:

split -b 10M content.txt splitted-content.txt-

This creates files approximately 10MB each. Useful for handling large files before transfer or parallel processing.

You can also use units like G for gigabytes or K for kilobytes:

split -b 1G largefile.bin chunk-

Splitting into a specific number of chunks

To divide a file into exactly N equal parts regardless of size, use -n:

split -n 4 content.txt splitted-content.txt-

This creates four roughly equal-sized files. With -n l/N, you can also split based on line count distribution:

split -n l/10 content.txt chunk-

Practical examples

Split a large CSV file for parallel processing:

split -l 50000 data.csv chunk-

for file in chunk-*; do
  process_csv "$file" &
done
wait

Split and compress output on the fly:

split -l 1024 content.txt - | gzip > content.txt.split.gz

Verify line counts after splitting:

for file in splitted-content.txt-*; do
  echo "$file: $(wc -l < "$file") lines"
done

Check total file size of chunks:

du -sh splitted-content.txt-* | tail -1

Reassembling split files

Combine chunks back into a single file:

cat splitted-content.txt-* > content.txt

With numeric suffixes, ensure proper ordering using brace expansion:

cat splitted-content.txt-{00..99} > content.txt

For alphabetic suffixes, the glob expansion typically respects the original order in modern shells, but verify the chunk count matches your expectation.

Handling permissions and ownership

The split command creates output files with default permissions (usually 0644). If you need specific permissions:

split -l 1024 content.txt splitted-content.txt-
chmod 600 splitted-content.txt-*
chown user:group splitted-content.txt-*

Common gotchas

Uneven final chunk: If file size isn’t evenly divisible, the last chunk contains fewer lines than specified. This is expected behavior and not an error.

Memory with very long lines: The split command processes line-by-line rather than in fixed buffers. Files with extremely long lines (>1GB per line) may cause memory issues, though this is rare in practice.

Glob expansion order: When reassembling with cat splitted-*, verify alphabetic or numeric order matches your expectation. Use explicit ranges or pipe through sort if unsure:

cat $(ls -1 splitted-content.txt-* | sort) > content.txt

Preserving content after split: Always verify the output before deleting the original file:

split -l 50000 large.txt chunk-
wc -l large.txt chunk-* | tail -1  # Verify line counts match
rm large.txt

The split command remains the most efficient and portable way to divide text files on Linux, requiring no additional tools or dependencies.