daff

Data diff tool for comparing tables and CSV files

TLDR

Compare two CSV files and show differences

$ daff [file1.csv] [file2.csv]

Compare with colored output

$ daff --color [file1.csv] [file2.csv]

Save diff output to a file

$ daff --output [changes.csv] [file1.csv] [file2.csv]

Apply a patch to update a file

$ daff patch [file.csv] [changes.csv]

Apply a patch in place

$ daff patch --inplace [file.csv] [changes.csv]

Three-way merge with a common parent

$ daff merge [parent.csv] [local.csv] [remote.csv]

Render a diff as HTML

$ daff render --output [diff.html] [diff.csv]

Copy/convert between formats

$ daff copy [input.csv] [output.tsv]

SYNOPSIS

daff [command] [options] [files]

daff is a data diffing tool designed for comparing tabular data such as CSV files, database tables, and spreadsheets. It provides git-like diff and patch operations specifically optimized for structured data formats.The tool can detect added, removed, and modified rows and columns, represent differences in various formats including HTML and colored terminal output, and apply patches to update data files. It's particularly useful for tracking changes in data exports, database dumps, and collaborative data editing workflows.

OPTIONS

--color

Force colored output (default: auto-detect)

--no-color

Disable colored output

--output FILE

Write output to file instead of stdout.

--input-format FORMAT

Input format (e.g., sqlite for comparing databases).

--www

Open diff in web browser.

--inplace

Modify input file directly (for patch and merge).

--unordered

Treat rows as unordered during comparison.

-h, --help

Display help and exit.

COMMANDS

a.csv b.csv (default)

Compare two tables and output differences.

patch [--inplace] a.csv patch.csv

Apply a patch file to a table.

merge [--inplace] parent.csv a.csv b.csv

Perform a three-way merge of tables.

render diff.csv

Render a diff as HTML.

trim source.csv

Trim whitespace from a table.

copy in.csv out.tsv

Copy and convert between table formats.

git

Use as a git diff driver for tabular files.

version

Show daff version.

CAVEATS

Data types are inferred from content; explicit type declarations are not supported. Very large files may require significant memory. Comparison of unordered data requires more computational resources. Binary data and special characters may not be handled correctly in all output formats.

HISTORY

daff was originally created by Paul Fitzpatrick as a tool for data versioning and collaboration. The name comes from "data diff." It has been implemented in multiple languages including JavaScript, Python, and Java. The tool was designed to bring version control concepts to data files, making it easier to track changes in datasets over time.