csvops

A fast CSV toolkit shipped three ways:

• 🖥️ CLI — single static Go binary for scripts and pipelines.
• 📦 Library — github.com/maherelgamil/csvops/pkg/csvops for embedding in Go programs.
• 🪟 Desktop app — paginated table view + one-click operations, built with Wails.

All three share the same engine: streaming CSV ops with progress callbacks and context.Context cancellation.

csvops split    --input big.csv     --rows 10000 --output-dir ./parts
csvops dedupe   --input users.csv   --output clean.csv --key email
csvops filter   --input users.csv   --column country --eq Egypt
csvops stats    --input data.csv
csvops preview  --input data.csv    --rows 10
csvops merge    --input-dir ./parts --output all.csv
csvops to-sqlite --input data.csv   --output data.db

Install (CLI)

Homebrew

brew install maherelgamil/tap/csvops

From a release binary

Grab the archive for your OS/arch from Releases and put csvops on your PATH.

From source

git clone https://github.com/maherelgamil/csvops.git
cd csvops
go build -o csvops .

Desktop app

Download the latest from Releases (look for tags starting with desktop-v):

• csvops-desktop-macos-universal.zip — macOS arm64 + amd64 fat binary
• csvops-desktop-windows-amd64.zip — Windows x64

macOS first launch: the app is currently ad-hoc signed, so Gatekeeper will block it. Open System Settings → Privacy & Security, scroll down, and click "Open Anyway" next to the warning. Apple Developer ID signing is on the roadmap.

What it does:

• Open or drag a CSV onto the window.
• Browse the whole file in a paginated table (50–1000 rows/page).
• Click any column header for instant per-column stats (unique, empty, top values).
• Run Filter / Dedupe / Split / Export to SQLite / Merge from the Actions menu — output is auto-suggested next to the input. Success banners include a Reveal-in-Finder button.

Build it yourself: see desktop/README.md.

Library

Use the same CSV engine in your own Go program:

import "github.com/maherelgamil/csvops/pkg/csvops"

ctx := context.Background()
res, err := csvops.Filter(ctx, csvops.FilterOptions{
    Input:      "users.csv",
    Output:     os.Stdout,
    Column:     "country",
    Eq:         strPtr("Egypt"),
    WithHeader: true,
    Progress:   func(done, total int64) { /* update UI */ },
})

Each operation (Split, Dedupe, Filter, Merge, Stats, Preview, ToSQLite) takes a typed Options struct and returns a typed Result. See pkg/csvops/ and the test files for full examples.

Commands

Command	Purpose
split	Split a large CSV into smaller chunks
merge	Combine all CSV files in a directory into one
dedupe	Remove duplicate rows by one or more key columns
filter	Keep rows matching eq / contains / gt / lt
stats	Row counts, unique values, empty cells, top values
preview	Pretty-print the first N rows as a table
to-sqlite	Import a CSV into a SQLite database

Run csvops <command> --help for the full flag list, or see docs/commands/.

split

Streams the input file and writes chunks of --rows lines to --output-dir.

csvops split --input big.csv --rows 10000 --output-dir ./parts --with-header --delimiter ","

--delimiter accepts a single character or \t for tab.

merge

Reads every *.csv file in --input-dir (sorted by name) and streams them into one output file. The header from the first file is written once when --with-header is set.

csvops merge --input-dir ./parts --output merged.csv

dedupe

csvops dedupe --input users.csv --output clean.csv --key email
csvops dedupe --input users.csv --output clean.csv --key first_name,last_name --case-sensitive
csvops dedupe --input users.csv --output clean.csv --key email --keep-last

• Output preserves the original file row order.
• Case-insensitive by default; pass --case-sensitive to compare exactly.
• --keep-last retains the last occurrence (default keeps the first).

filter

csvops filter --input users.csv --column country --eq Egypt
csvops filter --input users.csv --column name    --contains "ali"
csvops filter --input scores.csv --column score  --gt 80 --lt 100 --all

• Conditions combine with OR by default; pass --all for AND.
• A flag is only applied when explicitly set, so --eq="" matches empty values.
• Writes to --output if provided, otherwise to stdout.

stats

csvops stats --input data.csv
csvops stats --input data.csv --max-unique 5000

Prints row/column counts and a per-column table with unique value count, empty cell count, and top 3 values. --max-unique (default 100000) bounds memory on high-cardinality columns; columns that hit the cap are reported as >=N (capped).

preview

csvops preview --input data.csv --rows 20
csvops preview --input data.csv --no-header

to-sqlite

csvops to-sqlite --input data.csv --output data.db
csvops to-sqlite --input data.csv --output data.db --table users --if-exists append

• Pure-Go SQLite (modernc.org/sqlite), no CGO required.
• All columns created as TEXT. SQL identifiers are quoted, so column names and table names with spaces or special characters are safe.
• Default table name is derived from the input filename.
• --if-exists modes: replace (default, drops then re-creates), append (insert into existing), skip (no-op if table exists), fail (error if table exists).

Repo layout

cmd/                CLI commands (Cobra) — thin wrappers over pkg/csvops
pkg/csvops/         The CSV engine: Split, Dedupe, Filter, Merge, Stats, Preview, ToSQLite
desktop/            Wails React+TS desktop app, imports pkg/csvops
docs/commands/      Per-command CLI documentation

Contributing

PRs welcome — see CONTRIBUTING.md.

go build -o csvops .
go fmt ./...
go vet ./...
go test ./...

License

MIT © Maher El Gamil