zdiff

What it does

zdiff compares two files — at least one of which is gzip-compressed — by decompressing them on the fly and passing the result to diff. It shows a full line-by-line diff of the content, exactly like running diff on the decompressed files would, but without ever writing an uncompressed copy to disk. Every option diff understands can be passed through zdiff as well.

Beginner examples

• zdiff file1.gz file2.gz — show a line-by-line diff between two compressed files
• zdiff file1.gz file2 — compare a compressed file against a plain one
• zdiff -q file1.gz file2.gz — just report whether they differ, no details
• zdiff -u file1.gz file2.gz — unified diff format, like diff -u

zdiff app.conf.gz app.conf.new.gz

Advanced examples

• Review exactly what changed between two nightly compressed config snapshots: zdiff -u yesterday-config.gz today-config.gz
• Ignore incidental whitespace churn while reviewing a diff of a compressed export: zdiff -w old-export.csv.gz new-export.csv.gz
• Quickly gate a script on whether two compressed archives changed at all before doing expensive work: zdiff -q a.gz b.gz || echo "changed, reprocessing"
• Pipe unified-diff output into a patch-aware reviewer or save it for later: zdiff -u v1.log.gz v2.log.gz > changes.patch

zdiff -u /backups/etc-2026-07-04.tar.gz /backups/etc-2026-07-05.tar.gz | less

Common mistakes

• Expecting zdiff to compare the raw compressed bytes — it decompresses first, which is almost always what you want, but can surprise someone expecting a binary-level comparison (use cmp on the raw .gz files for that instead).
• Using zdiff on huge files and being overwhelmed by output — pair with -q first to check if there's even a difference before generating a full diff.
• Forgetting -i/-w exist and manually pre-processing files to strip case or whitespace differences before comparing.
• Confusing zdiff's job with zcmp's — reaching for zdiff when you only need a fast yes/no answer wastes time generating a diff nobody reads.

Tips

• Start with zdiff -q for a cheap changed/unchanged check, then re-run without -q (or with -u) only if you need to see specifics.
• Use -u when you intend to share or archive the diff — unified format is what most tooling (patch, code review tools) expects.
• Combine with -w/-i when comparing generated or formatted files where whitespace or case changes are expected noise, not meaningful differences.

Best practices

• Prefer zdiff over manually decompressing two files into temp copies just to run diff — it's fewer steps and leaves no cleanup behind.
• In config-management or backup-verification scripts, use zdiff -q for gating logic and reserve full diff output for human review or logging.
• Version-control the human-readable output of important zdiff -u comparisons (e.g., during audits) rather than the raw compressed files, since the diff is what's actually reviewed.

Real-world use cases

• Reviewing what changed between two compressed configuration backups before rolling a change forward or back.
• Auditing differences between two versions of a compressed log or data export during root-cause analysis of a data pipeline discrepancy.
• Verifying that a restored, re-compressed backup is functionally identical in content to the original, even if the compressed bytes differ due to a different gzip version.

Common interview questions

• What does zdiff do differently from diff? It transparently decompresses gzip-compressed inputs before comparing, so you never need to manually decompress files just to diff their contents.
• When would you use zcmp instead of zdiff? When you only need to know if two files differ (and where the first difference is), not what the actual line-by-line changes are — zcmp is faster and simpler for that.
• How would you get a unified diff between two compressed log files? zdiff -u file1.log.gz file2.log.gz.